The Planning Fallacy Cognitive.pdf

C H A P T E R O N E
The Planning Fallacy: Cognitive,
Motivational, and Social Origins
Roger Buehler,* Dale Griffin,†
and Johanna Peetz*
Contents
1. Defining the Planning Fallacy 2
2. Documenting the Planning Fallacy 4
3. Explaining the Planning Fallacy: The Original Cognitive Model 17
3.1. The inside versus outside view 17
3.2. Obstacles to using past experiences 18
3.3. Optimistic plans 19
4. Empirical Support for the Inside–Outside Model 20
5. Extending the Planning Fallacy: An Extended Inside–Outside Model 24
5.1. Identifying key elements of the inside focus 24
5.2. Effects of motivation: The mediating role of focus on plans 27
5.3. Perspective(s) and the planning fallacy 31
5.4. Social forces: Group processes accentuate plan focus 42
5.5. The behavioral impact of plans and predictions 46
5.6. Implications for debiasing 53
6. Concluding Perspectives 55
Acknowledgments 56
References 56
Abstract
The planning fallacy refers to a prediction phenomenon, all too familiar to many,
wherein people underestimate the time it will take to complete a future task,
despite knowledge that previous tasks have generally taken longer than
planned. In this chapter, we review theory and research on the planning fallacy,
with an emphasis on a programmatic series of investigations that we have
conducted on this topic. We first outline a definition of the planning fallacy,
explicate controversies and complexities surrounding its definition, and
summarize empirical research documenting the scope and generality of the
phenomenon. We then explore the origins of the planning fallacy, beginning
with the classic inside–outside cognitive model developed by Kahneman and
Advances in Experimental Social Psychology, Volume 43 # 2010 Elsevier Inc.
ISSN 0065-2601, DOI: 10.1016/S0065-2601(10)43001-4 All rights reserved.
* Psychology Department, Wilfrid Laurier University, Waterloo, Ontario, Canada
{
Sauder School of Business, University of British Columbia, Vancouver, BC, Canada
1

Tversky [Kahneman, D., & Tversky, A. (1979). Intuitive prediction: biases and
corrective procedures. TIMS Studies in Management Science, 12, 313–327].
Finally, we develop an extended inside–outside model that integrates empirical
research examining cognitive, motivational, social, and behavioral processes
underlying the planning fallacy.
Why do academic papers take longer to write than expected? Why do
home renovation projects take longer and cost more than expected? Why
do major public construction projects take longer and cost more than
expected, even though their scope and innovative features are typically
reduced along the way? The phenomenon of optimistic time predictions
undoubtedly comes in many forms and has many causes. Through a series of
programmatic investigations, we have been exploring one particular form of
optimistic time prediction that Kahneman and Tversky (1979) originally
termed the planning fallacy. In this chapter, we first outline the basic definition
of the planning fallacy, explicate controversies and complexities surrounding
its definition, summarize evidence supporting the classic inside–outside cog-
nitive model of Kahneman and Tversky (1979) and Kahneman and Lovallo
(1993), and then describe recent work that enriches the classic cognitive
model with motivational, social, and behavioral extensions.
1. Defining the Planning Fallacy
The planning fallacy refers to a readily observable phenomenon: the
conviction that a current project will go as well as planned even though
most projects from a relevant comparison set have failed to fulfill their
planned outcomes. The term was first introduced to the psychological
literature by Kahneman and Tversky (1979, 1982a, p. 415) to describe
people’s tendency ‘‘to underestimate the time required to complete a
project, even when they have considerable experience of past failures to
live up to planned schedules.’’
Kahneman and Tversky did not conduct empirical studies of the
planning fallacy, but offered a single case study that still stands as a defining
test case for theory and research on this phenomenon.
In 1976, one of us was involved in a project to develop a curriculum for a
new subject area for high schools in Israel. . . . Everyone on the team was
asked to write on a slip of paper the number of months that would be
needed to finish the project defined as having a complete report ready for
submission to the Ministry of Education. The estimates ranged from 18 to
30 months.
One of the team members—a distinguished expert in curriculum develop-
ment—was then posed a challenge by another team member: ‘‘Surely,
we’re not the only team to have tried to develop a curriculum where
2 Roger Buehler et al.

none existed before. Try to recall as many such projects as you can. Think
of them as they were in a stage comparable to ours at present. How long did
it take them at that point to reach completion?’’ After a long silence, the
curriculum expert said, with some discomfort, ‘‘First, I should say that not
all the teams that I can think of, that were at a comparable stage, ever did
complete their task. About 40% of them eventually gave up. Of the
remaining, I cannot think of any that completed their task in less than
seven years, nor of any that took more than ten.’’ He was then asked if he
had reason to believe that the present team was more skilled in curriculum
development than the earlier ones had been. ‘‘No,’’ he replied, ‘‘I cannot
think of any relevant factor that distinguishes us favorably from the teams I
have been thinking about. Indeed, my impression is that we are slightly
below average in terms of resources and potential. (Lovallo & Kahneman,
2003, p. 61; see also Kahneman & Tversky, 1982b)
Note that in this example so far, the narrator has assembled the ingre-
dients of the planning fallacy—a set of optimistic estimates for the current
task juxtaposed with the knowledge of a rather more pessimistic history of
such tasks—but the fallacy itself did not occur until the members of the team
ignored or neglected the lessons of the past. In this case, the planning fallacy
occurred when ‘‘. . .the members ignored the pessimistic information and
proceeded with the project. They finally completed the initiative 8 years
later, and their efforts went largely for naught—the resulting curriculum
was rarely used.’’ (Lovallo & Kahneman, 2003, p. 61)
The signature of the planning fallacy, then, is not that planners are
optimistic but that they maintain their optimism about the current project
in the face of historical evidence to the contrary. So, for example, even
though Denver’s ambitious Stapleton international airport project opened
16 months later than planned at a cost of at least $2 billion over budget, it
can only be defined as a case of the planning fallacy if the planners had
knowledge of a set of similar projects that had taken longer than planned
(which they surely did). Similarly, the fact that NASA’s most recent space
project has shown marked overruns in time and cost is not enough to define
it as a case of the planning fallacy without the knowledge (as summarized in
a recent U.S. General Accounting Office report) that 10 of the 13 preceding
projects had also been marked by time and cost overruns. A frustrated
project manager captures this point clearly: ‘‘A recent example I’ve had
was a task that the programmer promised he was going to finish it first in
2 days, I told him to be overly pessimistic, he said 2 weeks. The task was
scheduled for 4 [weeks]. Now, after 5 weeks, the programmer says he’s
2 days away from finishing.’’ With the planning fallacy, the future continues
to look rosier than the past, even as the future becomes the past.
Of course people can also underestimate the time required to complete a
task in contexts where they have no access to relevant historical base rates,
and there is value in understanding the causes and consequences of this more
The Planning Fallacy 3

general ‘‘optimistic bias’’ or ‘‘underestimation bias.’’ However, we believe
it is the remarkable combination of realistic knowledge about the past and
continuing optimism about the future that gives the planning fallacy its bite.
Definitions that do not acknowledge the paradoxical combination of opti-
mism about the future with realism about the past miss much of the richness
of the phenomenon. We were drawn to studying the planning fallacy
because in these types of predictions people do not forget history but
because people confidently make predictions that go against the history that
they know and remember. The study of the planning fallacy digs into the
deep psychology of George Bernard Shaw’s claim that ‘‘We learn from
history that we learn nothing from history.’’ (Shaw, 1903/1948, p. 485)
The lessons about when people do—and do not—learn from history also
make the study of the planning fallacy extremely relevant for social psychol-
ogists. The challenge of integrating memories of the past into our under-
standing of the present and our expectations for the future is a central
human problem that we face in every domain of our social lives. Studying
the planning fallacy provides insights about how hope, fear, disappointment,
and reason combine to shape not only time predictions but also many other
expectations for the future.
2. Documenting the Planning Fallacy
To identify examples of the planning fallacy requires two related but
separable findings. First, predictions of current task completion times must
be more optimistic than beliefs about the distribution of past completion
times for similar projects; second, predictions of current task completion
times must be more optimistic than actual outcomes. A high degree of
confidence is also diagnostic, but is not required by the classic definition.
The Standish Group publishes an annual survey about the success of
information technology projects in the United States; in 2009, they classi-
fied 32% of surveyed projects as successes, meaning they were delivered on
time, on budget, and with required features and functions; 44% were
challenged, which meant that they were delivered late, over budget, or
with less than required features; and 24% failed and were canceled or never
used (The Standish Group, 2009). In an earlier study (The Standish Group,
1995), the average time overrun was 222%, with 47% of all projects
reporting at least a 100% overrun. In a review of software estimation studies,
Molokken and Jrgensen (2003) report that between 65% and 85% of
software projects are completed after the predicted date; furthermore,
poor estimation was noted as the most common reason for IT project failure
(Nelson, 2007). In a comprehensive study of private sector planning and
forecasting, Norris (1971) compiled forecasted and completed times and

costs for close to 300 industrial research and design (RD) projects (these
300 projects were taken from a larger pool of 475 projects—the balance of
the projects were abandoned or changed. Notably, projects were generally
abandoned after the estimated point of completion). He found that projects
were, on average, about 250% over time budget and about 125% over cost
budget. In a more detailed study of 50 RD projects, Wilkes and Norris
(1972) found that substage estimates were just as optimistic as whole-project
estimates. Furthermore, there is ample historical evidence that cost and time
overruns have been characteristic of public works projects in both the
United States (Engerman Sokoloff, 2004) and Europe (Flyvbjerg et al.,
2005) for hundreds of years.
The tendency for project time estimation to show marked optimism
seems to hold for projects in the private and public sectors, as well as for
projects estimated to take from a few days to years to complete, and with
budgets from a few thousand to millions of dollars. However, there are two
major barriers to interpreting these applied real-world data in terms of the
planning fallacy. First, we have no direct measures of planners’ beliefs about
the distribution of relevant past outcomes. Although many surveys indicate
cost and time overruns are considered to be endemic in the high-tech and
project-management world, we are lacking the data that would tie beliefs
about the past to predictions about the future.
Second, the apparently optimistic time estimates for commercial projects
may represent duplicitous attempts to manipulate an audience for political
ends rather than true predictions about the future. Flyvbjerg et al. (2003,
2004, 2005) have argued that large public works are plagued by estimates
that are deliberately deceptive, either because politicians wish to develop
major projects but know that their true cost will not be borne by the
taxpayer, or because contractors lowball the estimated cost or effort predic-
tion to gain the contract, or because employees underestimate the cost and
scope of projects to gain political favor and project approval from their
managers. Flyvbjerg and colleagues provide anecdotal reports indicating
that employees are at least sometimes implicitly or explicitly rewarded for
deliberately underestimating project magnitude. Flyvbjerg also argues that
because the history of project overruns is so well-known, the continuing
pattern of optimistic forecasts in large-scale public projects has no other
reasonable explanation other than fraud and deception. Importantly, this
line of argument denies that people can honestly hold simultaneous beliefs
in a pessimistic past and an optimistic future. Whatever the role of the
‘‘honest’’ planning fallacy in project overruns, there is undoubtedly a strong
case for the role of deception and corruption in the underestimation of time
and cost in public works and private industry projects.
Thus, despite the applied importance of studying industrial and trans-
portation projects, and the plentiful data on the magnitude and generality of
optimistic biases in this domain, the cleanest laboratory for the study of the

planning fallacy seems to lie in the study of everyday personal projects.
Consider one familiar example: Academics who carry home a stuffed
briefcase full of work on Fridays, fully (and confidently) intending to
complete every task, are often aware that they have never gone beyond
the first one or two jobs on any previous weekend. Once again, this
illustrates the ability of people to hold two seemingly contradictory beliefs.
Although aware that most of their previous predictions were overly opti-
mistic, they believe (so it seems) that their current forecasts are realistic.
This phenomenon is universally recognized by academics, and so we
have often put it to the test with university students. However, the transla-
tion and operationalization from long-term projects is not as straightforward
as it might seem. For example, is the appropriate measure the duration of
time that students spend on the task, or is it the temporal distance to the
estimated completion time? We know from the literature on software
development projects that overestimation is typically found when measured
either way: as programmer hours or by the time to the completion of the
project. Nonetheless, larger overruns are generally found in estimates of
completion time (van Oorschot et al., 2002). Kahneman and Tversky’s
classic planning fallacy example was based on completion time: the curricu-
lum team estimated their completion date but there is no report of the team
jointly estimating how many hours they would spend on the task itself.
Moreover, we suspect that most project teams and most students measure
success in task completion by whether the product is delivered on time, and
rarely look back to assess whether the task took surprisingly many hours or
days of actual working time. Thus, in most of the studies we report here, the
key outcome measure is completion time rather than time-on-task dura-
tion. Moreover, in most cases, we examine target tasks that are of sufficient
magnitude that they require at least a full day of elapsed time between
prediction and completion.
A second decision we have faced is whether we should include or
exclude academic or personal tasks with a firm deadline. Strictly enforced
deadlines minimize the scope of the planning fallacy as they provide a tight
upper bound to the possible optimistic bias. Indeed, Tversky and Shafir
(1992) offered students $5 for answering and returning a questionnaire
within 5 days, 3 weeks, or with no deadline and observed return rates of
60%, 42%, and 25%, respectively, underlining the power of deadlines to
enforce timely completion. Similarly, Silk (2010) found that consumers
were much more likely to return mail-in rebates if given a 1-day or
1-week deadline rather than a more conventional 3-week deadline—even
though the consumers preferred the longer deadline if given a choice.
Furthermore, externally imposed deadlines are more effective in ensuring
timely task completion than are self-imposed deadlines (Ariely
Wertenbroch, 2002). Thus, our decision to include projects with deadlines—
and our general focus on academic projects with strict deadlines—undoubtedly

has led to a conservative estimate of the magnitude of the planning fallacy in
personal projects.
In one of our first assessments of the planning fallacy, we examined
predictions for completion of a computer-based tutorial session. Students
reported that they usually completed similar academic tasks about a day
before the deadline; when asked to predict when they would complete the
current tutorial, students estimated that they would be finished, on average,
about 6 days in advance of the deadline (Buehler et al., 1994). In fact, and
consistent with our two-point test for the planning fallacy, only about 30%
completed by their predicted date—most finished, as usual, considerably
closer to the deadline than they anticipated. We have obtained further
evidence that student participants firmly believe in their forecasts by asses-
sing their confidence (Buehler et al., 1994, Studies 2 and 3; Buehler et al.,
1995). First, participants reported how certain they were that they would
finish by the time they had predicted, on a scale ranging from 0% (not at all
certain) to 100% (completely certain). The 50% or halfway mark on this
scale indicated that respondents felt that it was a ‘‘toss-up’’ whether they
would finish by the time they predicted. For the most part, participants were
quite confident that they would meet their predictions: the average cer-
tainty rating was 74% for school assignments in one study, 84% in a second
study, and 70% for household tasks. In each case, approximately 40% of the
participants actually finished by the predicted time. It is notable that, in a
class survey, students reported that they typically completed about 32% of
their projects by their expected completion date, and were later than expected
about 68% of the time. The belief in (or knowledge of ) a ‘‘pessimistic past’’
seems to be widespread in this domain.
Figure 1.1 presents data from a related study (Griffin Buehler, 1999,
Study 1) where students reported probability estimates for their best guess
completion times for 10 current projects across their academic and personal
lives. Once again, about 45% of the projects were completed, compared to
an average confidence level of 73%. The calibration plot reveals two
interesting patterns: First, probability judgments are overly optimistic across
the entire range; this does not match the canonical pattern of overconfi-
dence in knowledge which is marked by underconfidence or underestima-
tion at the low end of the probability scale and overconfidence or
overestimation at the upper end, a pattern known as over-extremity
(Griffin Brenner, 2004). Thus, the phenomenon to explain is an opti-
mistic bias, not an extremity bias. Second, the slope between confidence
and likelihood of timely task completion is moderately steep, indicating that
in a relative sense, the predictions contain considerable information about
completion times.
As well as assessing judges’ confidence in their ‘‘best guess’’ predictions,
we have also employed a more demanding ‘‘fractile procedure’’
(Lichtenstein et al., 1982; Yates, 1990): In addition to their best guess

predictions, participants indicated times by which they were 50% certain,
75% certain, and 99% certain that they would finish a project (Buehler et al.,
1995). As seen in Fig. 1.2, the fractile measure revealed rampant optimistic
overconfidence: Only 13% of the participants finished their academic
projects by the time they reported as their 50% probability level, 19%
finished by the time of their 75% probability level, and 45% finished by
the time of their 99% probability level. The results for the 99% probability
level are especially striking: Even when asked to provide a prediction that
they feel virtually certain they will fulfill, students’ confidence far exceeds
their accomplishments.
We also documented a more dramatic example of optimistic time com-
pletion estimates by studying senior honors theses, a rare academic project
without a binding deadline. We waited until most students were approach-
ing the end of this year-long project, and then asked them to predict when
they realistically expected to submit the thesis, as well as when they would
submit it if ‘‘everything went as poorly as it possibly could.’’ The students’
realistic predictions were overly optimistic: Only 30% of the students fin-
ished the project by the predicted time. On average, the students took 55
days to complete their thesis, 22 days longer than they had anticipated, and
20
30
40
50
60
70
80
90
100
20 30 40 50 60 70 80 90 100
Estimated % complete
%
complete
0.12
0.32
0.19
0.17
0.17
Figure 1.1 Calibration curve for estimated and actual probability of meeting task
completion predictions. Numbers indicate the proportion of responses represented
by each point on the curve. Dashed line represents perfect calibration (Griffin
Buehler, 1999).

even 7 days longer than the average worst-case prediction. Interestingly,
despite the optimistic bias, the students’ forecasts were accurate in a relative
or ordinal sense: The predicted completion times were highly correlated
with actual completion times (r ¼ 0.77). Compared to others in the sample,
then, the students who predicted that they would take more time to finish
actually did take more time. Once again, we see that predictions can be
informative even in the presence of a marked optimistic bias.
A telephone survey of Canadian taxpayers revealed that the planning
fallacy is not limited to students and their school assignments (Buehler et al.,
1997). Those surveyed expected to mail their tax returns about a week
earlier than their typical date (3 weeks before the deadline rather than their
usual 2 weeks), but in accord with the planning fallacy, history repeated
itself: on average, reported typical behavior in the past accurately matched
present behavior (forms were completed in the usual 2 weeks), so that the
current predictions averaging 3 weeks before deadline were markedly
optimistic.
We should note, however, that current completion times do not always
match recollections of past behavior in our studies, nor does the definition
of the planning fallacy require such a match. Across many of our studies, we
find that the current task usually takes slightly longer than the recalled
typical time, implying either that the world is becoming progressively
more difficult or that people misremember the past in a somewhat positive
light. For example, in a study of holiday shopping predictions (Peetz et al.,
2010, Study 2), students reported that they typically completed their shop-
ping about 6.2 days before Christmas, and predicted they would complete
the current season’s shopping about 7.5 days before Christmas, but actually
completed it only 5.5 days before Christmas on average. Similarly, in studies
of student predictions for academic projects conducted in both Canada and
0
2
4
6
8
10
12
14
Best
estimate
50%
estimate
75%
estimate
99%
estimate
Actual
Days
37%
13%
19%
45%
Figure 1.2 Predicted completion times for academic tasks corresponding to specified
probability levels. Numbers above bars represent the percentage of projects finished by
predicted time (Buehler et al., 1995).

Japan (Buehler Griffin, 2003, Study 2; Buehler et al., 2003), we found
that predictions were more optimistic than the recalled date of a typical
project. Both Canadian and Japanese students predicted to finish about
2 days before the deadline and recalled that they typically finished about a
day before the deadline, but the current tasks were completed only half a
day before the deadline (see Fig. 1.3). In all cases, using recalled behavior to
make predictions would have been more accurate, at least on a mean level,
than using predicted behavior, but it would not have been perfect.
The evidence presented so far suggests that the tendency to underesti-
mate task completion times is a robust phenomenon. To characterize the
scope and generality of the prediction bias, we have also performed a more
exhaustive review of research outcomes from controlled studies of project
completion predictions and summarized these findings in Figs. 1.4 and 1.5
and associated Table 1.1. Our review builds upon previous summaries of the
literature (Roy et al., 2005) and includes additional studies. We included all
studies we could find that examined self-predictions (as opposed to observer
predictions) and reported untransformed measures of both predicted and
actual completion times. Because we were interested in characterizing
people’s natural ability to predict completion times, without assistance or
manipulation, we included only studies (or control conditions within stud-
ies) in which participants’ forecasts were not subject to experimental inter-
ventions designed to alter the information that predictors considered. In
addition, we classified the target tasks as ‘‘in-session’’ or ‘‘out-of-session’’
tasks based on whether participants completed the task during the research
0
1
2
3
4
5
6
7
Canada Japan
Days
before
deadline
Recalled typical
Predict
Actual
Figure 1.3 Recall of typical completion time, predicted completion time, and actual
completion time for a self-nominated school assignment in Canadian and Japanese
students (Buehler et al., 2003).

session (e.g., laboratory tasks) or completed it on their own time, outside the
research session. The dependent variables of interest were the predicted
and actual time to task completion, as opposed to the duration of time
spent working on the task. These two measures can (and usually do) diverge
markedly for out-of-session tasks (e.g., an essay with a 2-week deadline may
take 12 days from prediction to completion yet require only 5 h of actual time
on task), but are indistinguishable in the ‘‘in-session’’ tasks (e.g., a laboratory
task that takes 10 min of actual working time also has a time to completion of
10 min). We include a total of 65 comparisons of predicted and actual
outcomes, 35 from in-session tasks and 30 from out-of-session tasks.
Figure 1.4 presents a pair of histograms that summarizes the frequency of
studies showing underestimation bias versus overestimation bias separately
for in-session and out-of-session tasks. Both histograms are presented as
ratios (predicted time over actual time) to allow comparison between the
short in-session tasks (measured in minutes) and the longer, more complex
out-of-session tasks (measured in days). A ratio of 1 indicates a perfectly
accurate prediction, with ratios larger than 1 indicating overestimation (a
pessimistic belief that the task will take longer than it actually does) and
ratios less than 1 indicating underestimation (an optimistic belief that the
task will take less time than it actually does), consistent with the planning
fallacy. The contrast in the pattern of results is striking: for in-session studies,
where uncertainty is solely about the nature and difficulty of the target task,
there is considerable variability on both sides of the accuracy ratio, indicating
the presence of both overestimation and underestimation; for out-of-session
studies, where uncertainty is greater and comes not only from the nature of
the task but also from a host of other factors (e.g., competing tasks,
50
40
In-session tasks
30
Percent
of
studies
20
10
0
0 1 2
Ratio predicted to actual
completion time
3 4 5
50
40
Out-of-session tasks
30
Percent
of
studies
20
10
0
0.5
Ratio predicted to actual
completion time
0.75 1
Figure 1.4 Percent of studies finding underestimation (ratio 1) and overestimation
(ratio 1).

unsuspected difficulties in locating task-relevant materials, unexpected
interruptions, and self-control failure), underestimation is typical and over-
estimation is never observed.
Figure 1.5 further unpacks the difference between in-session and out-of-
session studies by plotting the same ratio of predicted to actual completion
time (in original ratio units on the Y-axis) against the actual task completion
time (in log units on the X-axis, to allow comparison between tasks taking
minutes to complete with tasks tasking weeks to complete). The top panel
presents the data from ‘‘in-session’’ studies, with actual times presented in
minutes, and the bottom panel presents the comparable data from ‘‘out-of-
session’’ studies, with actual times presented in days. For the in-session
studies, the degree and direction of bias is clearly related to the actual length
of the task, a relationship that has been documented previously (Roy
Christenfeld, 2008; Roy et al., 2005). Notably, for very short laboratory
tasks, overestimation is characteristic, whereas for somewhat longer tasks,
20s 15min
Actual task time
1 day 80 days
3
1
In-session tasks
0.3
Ratio
of
predicted
to
actual
completion
time
20s 15min
Actual task time
1 day 80 days
3
1
Out-of-session tasks
0.3
Ratio
of
predicted
to
actual
completion
time
Figure 1.5 Ratio of predicted to actual completion time as a function of actual
completion time.

Table 1.1 Summary of studies comparing predicted and actual task completion times
Reference Study Type Task Predicted Actual Unit
Buehler et al.
(1994)
1 2 Honors thesis [best guess predictions] 33.9 55.5 days
2 2 Self-nominated nonacademic project [partial
information group]
5 9.2 days
2 2 Self-nominated academic project [partial
information group]
5.8 10.7 days
3 2 Self-nominated school project 6.0 7.1 days
4 2 Computer tutorial assignment [control
group]
5.5 6.8 days
Buehler et al.
(1997)
1 2 Income tax forms [across refund/no refund
group]
22.2 14.0 days BDL
2 1 Anagram word puzzles [no-incentive
control group]
6.7 7 min
Buehler et al.
(2010b)
1 2 Self-nominated school project 11.4 13.7 days
Buehler and
Griffin (2003)
1 2 Christmas shopping [control group] 4.5 3.3 days BDL
2 2 Self-nominated school project [standard
control group]
1.5 0.3 days BDL
Buehler et al.
(2005)
1 2 School group project [final individual
predictions]
45.2 59.3 days
2 1 Puzzle task [final individual predictions] 13.1 13.8 min
3 2 Group take-home assignment [final
individual predictions]
1.9 2.5 days
Buehler et al.
(2003)
1 2 School assignment [Japanese sample] 1.7 0.3 days BDL
1 2 School assignment [Canadian sample] 1.8 0.7 days BDL
2 Summer research project 14.2 6.7 days BDL
(continued)

Table 1.1 (continued)
Burt and Kemp
(1994)
1 1 Bookshop 120 44 s
1 1 Cards 180 52 s
1 1 Library form 300 74 s
1 1 Library 133 139 s
1 1 Walk 600 249 s
2 1 Purchase 420 433 s
2 1 Balance sheet 900 557 s
2 1 Letter 600 611 s
2 1 Proofread 900 690 s
2 1 Library 840 734 s
Byram (1997) 1 1 Assemble computer stand [subset that
performed the task]
65.7 76.1 min
4 1 Origami task [final prediction, across
immediate/delay group]
6.6 9.7 min
5 1 Origami task [no-incentive control group] 10.1 9.8 min
Connolly and
Dean (1997)
1 2 Programming assignment #1 [predict whole
task first group]
8.3 8.7 h
1 2 Programming assignment #2 [predict whole
task first group]
9.9 13.4 h
Deslaurier (2002) 1 2 Single-submission essay task [control group] 7.72 10.2 days
1 2 Multiple-submission essay task [control
group]
8.85 10.4 days
Forsyth and Burt
(2008)
1 1 Office tasks A [scheduled as group] 17.9 12.8 min
1 1 Office tasks B [scheduled as group] 15 14.5 min
2 1 Office tasks A [scheduled as group] 22.4 71.6 min
2 1 Office tasks B [scheduled as group] 21.2 85.5 min
14

Francis-Smythe
and Robertson
(1999)
1 1 Spell-check 429 357 s
Hinds (1999) 1 1 Operate cell phone [novice sample, unaided
group]
15.7 31.5 min
Josephs and
Hahn (1995)
3 1 Read manuscript [short format, across slow/
fast readers]
35.8 65.0 min
3 1 Read manuscript [long format, across slow/
fast readers]
25.5 68.6 min
4 1 Read manuscript [short format, across slow/
fast readers]
55.5 69.8 min
4 1 Read manuscript [long format, across slow/
fast readers]
44.2 77.6 min
König (2005) pilot 1 Answer questions 49 59.9 min
Koole and Van’t
Spijker (2000)
1 2 Essay task [goal condition] 1.7 2.9 days
Kruger and
Evans (2004)
3 1 Format document [packed group] 17.3 39.7 min
4 1 Assemble food tray [packed group] 23.5 33.6 min
5 1 Format document [simple; packed group] 7.9 8.1 min
5 1 Format document [complex; packed group] 11.9 23.7 min
Newby-Clark
et al. (2000)
1 2 Major school assignment [realistic scenario;
final prediction]
1.1 0.3 days BDL
3 2 Major school assignment [no scenario
control]
2.0 0.8 days BDL
(continued)
15

Table 1.1 (continued)
Peetz et al.
(2010)
1 2 Major school assignment 2.5 1.3 days BDL
2 2 Christmas shopping [predicted in October] 9.2 5.7 days BDL
2 2 Christmas shopping [predicted in
December]
5.7 5.4 days BDL
5 2 Essay task [predicted 3 weeks BDL] 1.6 0.6 days BDL
5 2 Essay task [predicted 1 week BDL] 1.1 0.3 days BDL
Pezzo et al.
(2006a)
1 2 Read story 8.2 12 days
1 2 Proof-read lists 6.0 7.9 days
Roy and
Christenfeld
(2007)
1 1 Origami task [no prior experience group;
prediction group]
20 11 min
Sanna et al.
(2005)
1 2 Group research proposal [control group] 8.3 3.1 days BDL
2 1 Group desk assembly [control group] 42.4 50.0 min
3 1 Group desk assembly [control group] 42.3 53.2 min
Thomas and
Handley
(2008)
2 1 Build toy castle [no-anchor control group] 619 652 s
Thomas et al.
(2003)
1 1 Tower of Hanoi [no-experience control
group]
61.7 21.7 s
Zhang and
Fishbach
(2010)
5 2 GRE questions [estimate group; expect
simple task]
59.7 62.4 h
5 2 GRE questions [estimate group; expect
difficult task]
29.3 35.1 h
Note. Type of task refers to whether the task was completed as part of the experimental session (1 ¼ in-session, 2 ¼ out-of-session). For studies with experimental
manipulations, the control group selection is specified in brackets. BDL ¼ before deadline.

underestimation becomes common. This is consistent with a floor effect on
a bounded scale for these very short tasks. However, for out-of-session
studies, the degree of underestimation is consistent across the range of task
length, indicating that optimistic time predictions are characteristic of tasks
carried out in a real-world context, whether they are relatively short-term
or long-term projects. The starkly contrasting patterns between in-session
and out-of-session prediction tasks shown in both Figs. 1.4 and 1.5 provide
justification for our view that the planning fallacy is a phenomenon charac-
teristic of familiar tasks carried out under conditions of considerable uncer-
tainty—and should be studied under those conditions (Griffin Buehler,
2005). Furthermore, these distinct patterns cast doubt on the utility of
attempts to build a single account of prediction bias that applies readily to
novel and familiar tasks, short and long tasks, and tasks performed in and
outside of the laboratory (cf. Roy et al., 2005).
3. Explaining the Planning Fallacy: The Original
Cognitive Model
3.1. The inside versus outside view
Given the prevalence of optimistic predictions, and ample empirical evi-
dence of the planning fallacy, we now turn to examining the psychological
mechanisms that underlie people’s optimistic forecasts. In particular, how do
people segregate their general theories about their predictions (i.e., that they
are usually unrealistically optimistic) from their specific expectations for an
upcoming task? Kahneman and Tversky (1979) explained the prediction
failure of the curriculum development team through the inside versus outside
analysis of the planning fallacy. This analysis builds upon a perceptual
metaphor of how people view a planned project. In the curriculum devel-
opment example, the group of authors focused on the specific qualities of the
current task, and seemed to look inside their representation of the developing
project to assess its difficulty. The group of authors failed, however, to look
outside of the specific project to evaluate the relevant distribution of compa-
rable projects. Even when they asked for information about the outside
viewpoint, they neglected to incorporate it in their predictions or even to
moderate their confidence. An inside or internal view of a task focuses on
singular information: specific aspects of the target task that might lead to
longer or shorter completion times. An outside or external view of the task
focuses on distributional information: how the current task fits into the set of
related tasks. Thus, the two general approaches to prediction differ primarily
in whether individuals treat the target task as a unique case or as an instance of
a category or ensemble of similar problems.

To explain the planning fallacy using this model, we need to account for
two things: why people fail to incorporate the distributional or outside
information in their completion estimates, and why the resulting focus on
singular information through the inside perspective leads to overoptimism.
For a number of reasons, the singular or ‘‘narrow’’ frame seems to be most
natural when developing plans and predictions as well as making decisions
about the future (Kahneman Lovallo, 1993; Kahneman Tversky,
1979). People tend to generate their predictions by considering the unique
features of the task at hand, and constructing a scenario about their future
progress on that task. This is, by and large, what ‘‘planning’’ means to most
people: developing a series of steps that will lead from the beginning to a
successful conclusion of the project. In deciding how much work to take
home for the weekend, for example, individuals may try to imagine when
they will start a particular project and how many hours of actual working
time they will require. They may (or may not) also consider the other
activities they have planned for the weekend and try to determine precisely
when, where, and how they will find time to work on the project. They
may (or may not) even try to envision potential obstacles to completing the
project, and how these obstacles will be overcome. Essentially this inside or
internal approach to prediction involves sketching out a scenario that
captures how the specific future project is likely to unfold. Several theorists
have similarly noted the tendency to construct scenarios, narratives, or
mental simulations as a basis for predicting the future (Dawes, 1988;
Dougherty et al., 1997; Dunning, 2007; Griffin et al., 1990; Kahneman
Lovallo, 1993; Kahneman Tversky, 1982c).
3.2. Obstacles to using past experiences
Despite the predictive value of past experiences, our findings suggest that
people neglect them while forming predictions. Why might this be so? One
answer was suggested above: prediction, by its very nature, elicits a focus on
the future rather than the past, and this future orientation may prevent
individuals from looking backwards. However, a failure to use personal base
rates need not always result from inattention to the past. People may
sometimes attend to their past, but nevertheless fail to incorporate this
information into their predictions because it does not seem relevant. People
may have difficulty extracting an appropriate set of past experiences; the
various instances seem so different from each other that individuals cannot
compare them meaningfully (Kahneman Lovallo, 1993; Kahneman
Tversky, 1979). As a British rugby-player-turned-commentator noted
‘‘Each game is unique, and this one is no different from any other.’’
People may also interpret the past in a manner that reduces its pertinence
to the current prediction. The meaning and relevance of any past behavior
depends largely on an individual’s explanation of why it occurred. Whereas

certain types of attributions link a past event to the present and future, other
attributions isolate the past. To the extent that people perceive a previous
episode to be caused by external, unstable, and specific factors, they need
not connect its outcome to future occasions. For example, a woman may
attribute her inability to complete a past weekend task to an unexpected
visit by her former college roommate. Thus, she may generalize the previ-
ous failure only to weekends when that external and specific factor is
present. Knowing that her college friend is out of town, she may assume
she will be able to readily attain her objectives.
We suggest that people often make attributions that diminish the rele-
vance of past experiences to their current task. People are probably most
inclined to deny the significance of their personal history when they dislike
its implications (e.g., that a project will take longer than they hope). If they
are reminded of a past episode that could challenge their optimistic plans,
they may invoke attributions that render the experience uninformative for
the present forecast. This analysis is consistent with evidence that individuals
are inclined to explain away negative personal outcomes (for reviews, see
Miller Ross, 1975; Taylor Brown, 1988). People’s use of others’
experiences are presumably restricted by the same two factors: a focus on
the future reduces the salience of others’ experiences, and the tendency to
attribute others’ outcomes to their dispositions (Gilbert Malone, 1995)
limits the inferential value of others’ experiences. Furthermore, our under-
standing of other people’s experiences is typically associated with uncer-
tainty about what actually happened; consequently, we can readily cast
doubt on the generalizability of those experiences. To quote Douglas
Adams, ‘‘Human beings, who are almost unique in having the ability to
learn from the experience of others, are also remarkable for their apparent
disinclination to do so.’’ (Adams Carwardine, 1991, p. 116)
In sum, we note three particular impediments to using the outside
perspective in estimating task completion times: the forward nature of
prediction which elicits a focus on future scenarios, the elusive definition
of ‘‘similar’’ experiences, and attributional processes that diminish the
relevance of the past to the present.
3.3. Optimistic plans
People’s completion estimates are likely to be overly optimistic if their
forecasts are based exclusively on plan-based, future scenarios. A problem
with the scenario approach is that people generally fail to appreciate the vast
number of ways in which the future may unfold (Arkes et al., 1988;
Fischhoff et al., 1978; Hoch, 1985; Shaklee Fischhoff, 1982). For
instance, expert auto mechanics typically consider only a small subset of
the possible things that can go wrong with a car, and hence underestimate
the probability of a breakdown (Fischhoff et al., 1978). Similarly, when

individuals imagine the future, they often fail to entertain alternatives to
their favored scenario and do not consider the implications of the uncer-
tainty inherent in every detail of a constructed scenario (Griffin et al., 1990;
Hoch, 1985). When individuals are asked to predict based on ‘‘best guess’’
scenarios, their forecasts are generally indistinguishable from those generated
by ‘‘best-case’’ scenarios (Griffin et al., 1990; Newby-Clark et al., 2000).
The act of scenario construction itself may lead people to exaggerate the
likelihood of the scenario unfolding as envisioned. Individuals instructed to
imagine hypothetical outcomes for events ranging from football games to
presidential elections subsequently regard these imagined events as more
likely (for reviews, see Gregory Duran, 2001; Koehler, 1991). Focusing
on the target event (the successful completion of a set of plans) may lead a
predictor to ignore or underweight the chances that some other event will
occur. Even when a particular scenario is relatively probable, a priori,
chance will still usually favor the whole set of possible alternative events
because there are so many (Dawes, 1988; Kahneman Lovallo, 1993).
4. Empirical Support for the Inside–Outside
Model
An important goal of our initial studies of the planning fallacy was to
test the inside–outside account. We first looked for evidence that a focus on
the future—and a consequent neglect of the past—was indeed characteristic
of the planning process. In several studies, we included ‘‘think-aloud’’
procedures to record the ‘‘online’’ narratives of participants as they esti-
mated their completion times for various tasks, most of which had specific
deadlines. We instructed respondents to say aloud every thought or idea that
came to mind while predicting when they would finish an upcoming
project. We later analyzed the verbal protocols for evidence that people
focus on plan-based scenarios for the task at hand, rather than distributional
information such as their previous experiences. In a typical study of aca-
demic and home projects (Buehler et al., 1994, Study 3), the majority of
respondents’ thoughts were directed toward the future (M ¼ 74%). The
participants focused overwhelmingly on their plans for the current project,
for the most part describing scenarios in which they finished the task
without problems arising (only 3% of respondents’ thoughts included
potential impediments); their verbal protocols revealed an almost total
neglect of other kinds of information, including their own past experiences
(7% of thoughts) or others’ experiences with similar projects (1%). We have
replicated this pattern of results a number of times using written thought-
listing measures instead of think-aloud procedures, and collecting retrospec-
tive rather than online measures of thoughts. Thought-listings also revealed

a consistent neglect of another piece of information in favor of the focus on
future plans. In all studies, few students report thinking about the role of
deadlines when constructing their forecasts of completion times. For exam-
ple, in one study with varying deadlines (Buehler et al., 1994, Study 2),
students’ predictions were only weakly associated with their self-reported
deadlines (r ¼ 0.23, p 0.09), whereas their reported completion times
were strongly associated with the deadlines (r ¼ 0.82, p 0.001).
In one particularly interesting analysis of thought focus, we compared
the reported thought content of Canadian and Japanese students (Buehler
et al., 2003). Participants named a school assignment that was due in the
next 4 weeks, reported best guess predictions for that project, described
their thoughts while they made their predictions, reported when they
typically finished similar assignments in the past, and later reported when
they actually finished the target assignment. As described earlier, both
Japanese and Canadian participants showed the classic planning fallacy
dissociation between their current predictions and their reports of typical
completion times. Furthermore, as seen in Fig. 1.6, the thought-listings also
revealed a very similar pattern across cultures. Consistent with previous
findings, most participants (70%) focused on their plans for completing the
task, whereas only about one quarter of the participants (26%) referred to
their past experiences (which typically implied that a more conservative
prediction was in order), and very few participants contemplated the poten-
tial problems and delays they might encounter. The Canadian and Japanese
samples did not differ significantly in their use of any of the thought
categories.
0
10
20
30
40
50
60
70
80
90
100
Canada Japan
Percentage
of
respondents
Plans
Obstacles
Past
Disposition
Figure 1.6 Focus of thoughts in Canadian and Japanese students predicting their
completion for a self-nominated school assignment (Buehler et al., 2003).

In another experiment (Buehler et al., 1994, Study 4), we examined
whether the relative neglect of past experiences was due to a lack of
attention or due to the perceived irrelevance of the past. Participants
predicted when they would finish a standard, 1-h computer assignment.
Those in the Recall condition reported on their previous experiences with
similar assignments just before making their predictions. They indicated
how far before the deadlines they had typically finished school assignments
similar to the present one. Control participants were not asked any ques-
tions concerning their past experiences until after they had reported their
predictions. In both conditions, participants remembered finishing the
majority of their previous projects very close to the relevant deadlines.
Participants in the Recall condition—who had just focused on their usual
deadline-snatching pattern of completion times—nonetheless underesti-
mated their actual completion times just as much as those in the standard
Control condition. Even in the Recall condition, only 12% of participants
reported thinking about past experiences when making their current plans
(vs. 2% in the Control condition). Again, this pattern has been replicated in
other studies (e.g., Buehler Griffin, 2003), indicating that attention to and
awareness of the past is not enough to make the past relevant to the future.
The relevance account is further strengthened by a demonstration that
forcing the past to become relevant eliminates the optimistic prediction bias
for completion times. Our manipulation required participants to actively
link their past experiences with their specific plans for an upcoming task. In
this ‘‘recall-relevance’’ manipulation, participants first indicated the date
and time they would finish a computer assignment if they finished it as far
before its deadline as they typically completed assignments (Buehler et al.,
1994, Study 4). They were then required to describe a plausible scenario—
based on their past experiences—that would result in their completing the
computer assignment at their typical time. This procedure was designed to
prevent participants from either ignoring past experiences with similar tasks
or discounting the relevance of those experiences. After writing the hypo-
thetical scenario, participants made predictions for the computer assign-
ment. For this task, at least, the recall-relevance manipulation successfully
eliminated the usual optimistic bias.
Recall that a third reason that people may fail to incorporate the lessons
of past experiences into their future predictions is that people may diminish
the relevance of the past through attributional means. To test the role of
attributional processes in the planning fallacy, we have asked participants to
recall an occasion when they had failed to complete a task by the time they
had originally anticipated, and then to recall a similar prediction failure
experienced by a close acquaintance (Buehler et al., 1994, Study 3). Next,
we asked them to explain why each of the two tasks was not finished by the
expected time and coded them using the dimensions of the Attributional
Style Questionnaire (Peterson et al., 1982). The reasons participants reported

for their own lateness were more external, transitory, and specific than the
reasons they provided for similar tardiness by close acquaintances. Partici-
pants attributed their own lateness to such rare events as their computer
monitor ‘‘frying’’ while they were typing their final English essay, whereas
others’ failures seemed to reflect enduring personal problems with time
management. In a related study (Buehler et al., 1994, Study 4), participants
who either succeeded or failed in meeting their predictions rated their own
reasons for the success or failure. Those who finished late rated their reasons
as significantly more transitory and specific than those who finished on time.
Both studies reveal that people interpret their own tardiness in a manner that
makes it seem unique and unlikely to recur.
In the cross-cultural study mentioned previously (Buehler et al., 2003),
we asked participants to describe a past prediction failure of their own and
that of a school friend and to explain why each of these had occurred. Once
again, we found a moderate self-serving bias in Canadians’ attributions for
failures to complete tasks by the expected time, one that could serve to
diminish the relevance of past failures to future forecasts. Canadian students
reported causes that were more external, transitory, and specific than the
causes they reported for a friend’s late completion. However, we did not
expect that the same attributional bias would appear in Japanese students,
given that theorists propose Japanese culture is marked by social pressure
toward self-derogation rather than self-enhancement, at least publicly (e.g.,
Heine et al., 2000, 2001). Consistent with these cultural theories, the
Japanese students demonstrated a self-derogation bias in their attributions
for failing to complete tasks on time: they attributed their own prediction
failures to causes that were more internal, stable, and global than those of a
friend. This finding implies that the planning fallacy may be supported by
attributional processes to a lesser degree in Japanese than in Western
cultures, and raises the intriguing possibility that Japanese participants
might be more influenced by reminders of past failures than are Western
participants.
We have also validated the role of the perceived irrelevance of the past
(at least in Western cultures) through interviews with software engineers
engaged on major projects at a firm specializing in taxation software for
government use. Although this small sample does not provide a confirma-
tory test of the attributional hypothesis, it does provide vivid demonstra-
tions of the seemingly unassailable logic that drives the dissociation of past
from future projects. Of the dozen software engineers interviewed, there
was virtual unanimous agreement that past projects tended to be late and that
the results of past projects could not be used to inform future predictions.
Participants were first asked ‘‘When you made the prediction. . .did you
think of past projects and their outcomes?’’ Typical answers included
‘‘No. . .because it’s a unique working environment and I’ve never worked
on anything like it’’; ‘‘No, not relevant. It’s not the same kind of project at all.’’

Participants were also asked ‘‘When you made this prediction did you think
about other people’s projects here or within the field in general?’’ Answers
included: ‘‘There didn’t seem to be anything that was comparable. This is a bit
unique. A lot of it relies on new technology, new software on which to build
reports, etc. It’s difficult to compare to other projects’’; ‘‘No, not on this one.
The nature of this project is a little strange; it’s not like anything I’ve
ever worked on’’; ‘‘No comparison, because this type of thing hasn’t been
done before.’’
Overall, the evidence for the cognitive processes underlying the inside
model of prediction is highly consistent, whether in terms of thought-listing
analysis, manipulations of attention to and perceived relevance of past
experiences, or measures of attributional processes that diminish the rele-
vance of the past to future outcomes. In the next sections, we go beyond
documenting the processes underlying the classic model of the planning
fallacy, and use these processes to explain the effects of a broader range of
variables, resulting in an extended model (Fig. 1.7 presents a comparison of
the classic and extended inside–outside model).
5. Extending the Planning Fallacy: An Extended
Inside–Outside Model
5.1. Identifying key elements of the inside focus
A basic axiom of the planning fallacy is that the future is perceived to be
rosier than the past; realistically pessimistic lessons from the past fade from a
forecaster’s attention in light of optimistic plans about the future. However,
according to Kahneman and Tversky’s (1979) initial definition, the inside
view includes all case-specific content including possible problems or
obstacles to completion as well as plans for how to overcome them. Thus,
an inside view can, at least in theory, vary in the balance of its content
between plans for success and possible obstacles or interruptions. This led us
to two guiding hypotheses about the content of the inside view on a given
prediction: (1) the content will, in general, be biased toward a best-case
scenario dominated by plans for success; (2) the balance of the content
between plans for success and obstacles to completion is a continuous
variable that varies according to the psychological state of the forecaster
and influences the optimism of the resulting forecast. Thus, in the research
that follows, we extend the inside–outside view of the planning fallacy by
drawing a distinction not only between inside and outside perspectives, but
also between varying amounts of positive focus within the inside
perspective.
To evaluate the first of these hypotheses, Newby-Clark et al. (2000)
assessed the overall positivity of people’s scenarios about future projects by

contrasting thought-listings and predictions created by forecasters assuming
either realistic, worst-case, or best-case future scenarios. Notably, the con-
tent of the best-case scenarios was rated as only slightly (although signifi-
cantly) more optimistic than the content for the realistic scenarios, and the
number of future problems mentioned and the rated plausibility of the
scenarios did not differ between realistic and best-case conditions (which
both differed markedly from the worst-case scenario condition). A follow-
up study revealed that when predictors constructed multiple scenarios to
describe the time-course of a project, the first scenario was the most
optimistic in content, contained the fewest problems, and was rated as
most plausible. Furthermore, only the predicted completion time associated
with the first scenario was uniquely associated with the final ‘‘best guess’’
prediction. Taken together, these findings imply that planning a task and
estimating its completion is similar to—although not identical to—assuming
the best and working from there.
Extended inside–outside model
Classic inside–outside model
+
+
Inside view
(case-specific)
Outside view
(reference-class)
Prediction
optimism
Behavior
More use of memories
More use of base rates
More plan focus
Less obstacle focus
+
Closed versus
open tasks
More use of theories
-
+
+
-
-
Greater social power
Narrower perspective
Social pressure
Closer temporal
distance
+
+
-
First versus third-
person imagery
Actor versus
observer status
Incentives Motivation
Inside view
(case-specific)
Outside view
(reference-class)
Behavior
+
-
Prediction
optimism
Group versus
individual
Figure 1.7 Summary of the classic and extended inside–outside model of the planning
fallacy.

According to our second hypothesis, scenarios can vary in the extent to
which they are tightly focused on an optimistic future plan versus open to
acknowledging the possibility of obstacles, both foreseen and unforeseeable.
Following this logic, a greater than normal focus on detailed planning could
exacerbate the degree of optimistic bias. In two studies examining school
assignments and holiday shopping plans, we experimentally manipulated the
degree of focus on plans by having some participants make detailed, step-
by-step plans about a current project whereas other participants merely
reported their predictions about project completions (Buehler Griffin,
2003). In both studies, we instructed a subset of the participants, those in the
‘‘planning-focus’’ condition, to formulate highly specific plans for complet-
ing their target task. These participants considered and described in detail
when, where, and how they would carry out the task. This enhanced focus
on concrete plans produced highly optimistic predictions. In the ‘‘Christmas
shopping’’ study, for example, participants in the planning-focus condition
predicted they would complete their designated shopping list more than a
week before Christmas day (M ¼ 7.5 days) whereas control participants
expected to finish approximately 4 days before Christmas (M ¼ 4.5 days).
Follow-up interviews revealed that participants in both conditions finished
their designated shopping list only 3 days before Christmas. A similar pattern
was obtained for the study involving school assignments. Increasing the
focus on specific plans produced more optimistic forecasts because this
focus, induced at the time of prediction, substantially affected predictions
but did not affect actual completion times.
This pattern of effects may seem surprising in light of the apparent
similarity of the planning focus manipulation to manipulations of unpacking
(e.g., Kruger Evans, 2004) and implementation intentions (e.g.,
Gollwitzer, 1999). First, unpacking a complete project into subcomponent
tasks makes people less optimistic in their time predictions because the
subtasks serve as specific reminders of the many steps that need to be
completed, but may be neglected in an overall scenario-based project plan
(Kruger Evans, 2004). However, the effects of breaking down a task into
concrete details may depend on the type of information that is highlighted.
If people are induced to focus on otherwise ignored nonfocal aspects of a
task, as in the unpacking manipulation, this should lead to less optimism. If,
however, people are induced to focus in greater concrete detail on how
they will carry out the central or higher level task components (without
contemplating the otherwise ignored subcomponents of the task), then we
should find greater optimism. Given that the planning focus instructions
emphasized the creation of a flowing scenario for a unitary task (e.g.,
‘‘. . .Try to provide a complete picture, from beginning to end, of how
this assignment will be completed’’), it probably led participants to focus on
their plans for carrying out central task components, rather than to recognize
previously ignored aspects of the task.

Second, there is considerable research indicating that making specific
plans can have facilitating effects on behavior (e.g., Gollwitzer, 1999;
Gollwitzer Brandstätter, 1997; see also Armor Taylor, 2003; Koole
Van’t Spijker, 2000), which we do not find in these studies. Participants
who share a common goal intention (e.g., a desire to maintain an exercise
program or diet) are more likely to succeed if they are prompted to generate
implementation intentions that specify precisely when, where, and how
they will act to achieve the goal. The implementation intention is effective
because it links anticipated opportunities with effective goal-directed beha-
viors. A key variable controlling the impact of a concrete plan on actual
completion times, we suggest, is the nature of the barriers to project
completion. When the barriers to completion consist primarily of internal
(and therefore more controllable) factors such as forgetfulness, lack of
motivation, and weakness of will, we expect the effect of plans on behavior
to be substantial. However, when the barriers to completion consist pri-
marily of external (and therefore less controllable) factors such as competing
projects, unexpected problems in the task itself, or missing resources, we
expect that the effect on behavior will be minimal. We believe that the kind
of projects we have been studying—the completion of income tax forms,
major school assignments, and long-term personal projects—may fit the
latter pattern. These are long-term projects that require sustained effort over
several days or weeks, and the problems that are encountered are generally
not because of self-control or forgetfulness about goal completion. We will
return to the question of when optimistic planning processes influence
behavior in a later section.
5.2. Effects of motivation: The mediating role of
focus on plans
Our emphasis on the balance of positive versus negative content in the
extended inside–outside model also highlights the role of motivation in
enhancing or exaggerating a narrow focus on planning for success. Direc-
tional motivation in service of the desire to finish tasks quickly may lead to a
greater focus on future plans and a reduced focus on possible obstacles (e.g.,
problems, distractions, or other threats to completion), a form of motivated
reasoning (Kunda, 1990) or desirability bias (Krizan Windschitl, 2007).
We hypothesize that as the motivated individual relies on more focused and
biased plans for success, the resulting predictions should exhibit an even
greater optimistic prediction bias—assuming that completion behavior is
not equally influenced by the optimistic plans and associated goals. Our
contention is that although it is relatively easy and straightforward to
generate plans and predictions that correspond with one’s current goal, it
will usually be much more difficult to translate that goal into behavior.
People’s task completion behaviors occur long after the prediction is

generated and are subject to a wide range of external factors that can
intervene between prediction and completion behavior. To the extent
that a task is prone to influence from these external factors, people’s
incentive-induced goals at the time of prediction may have little, if any,
impact on actual completion time.
In an initial nonexperimental field test of the role of motivated
reasoning, we surveyed Canadian taxpayers and asked them to predict
when they would file their income tax forms (Buehler et al., 1997). The
study took advantage of a naturally occurring variation in people’s desire to
finish the task promptly. Individuals expecting a tax refund have an incen-
tive to submit the forms early (and thus receive an early refund). These
motivated respondents predicted that they would send in their tax returns
about 10 days earlier than nonmotivated respondents, but managed to send
them in only about 3 days earlier. As a result, the optimistic prediction bias
was markedly greater for the motivated respondents. Furthermore, the
individuals expecting a refund appeared to give less weight to their past
experiences in making their predictions: reports of past completion times
were less related to predicted completion times for participants who
expected a refund than for participants who did not.
This field study is consistent with the hypothesis that motivation
enhances the planning fallacy through a heightened focus on future plans.
However, it was a nonexperimental demonstration and subject to a host of
alternative explanations; in addition, the field setting prevented a detailed
measure of the hypothesized process. As a consequence, we conducted a
follow-up laboratory study that was tightly controlled and provided detailed
process data but was rather far removed from the defining real-world
phenomena. This experiment used a word-puzzle task to replicate the
incentive effect on prediction in an experimental setting and to measure
the mediating cognitive mechanisms. After taking part in a practice trial,
participants completed three trials with feedback (completion times ranged
from 5 to 7 min) so that they had prior experiences on which to base their
subsequent prediction. Some participants, those in the speed incentive
condition, were then promised monetary incentives if their completion
time in the next round was substantially (1 or 2 min) faster than their
previous trial time. As expected, monetary incentives for speedy completion
of the task led to more optimistically biased predictions, increased attention
to detailed future plans, reduced attention to—and perceived relevance
of—past experiences, and reduced thoughts about possible obstacles. In
other words, the speed incentives appeared to elicit the double pattern of
cognitive processes that constitute the planning fallacy—an exaggerated
focus on singular information (plan-based scenarios) at the expense of
relevant distributional information (past experiences), and an exaggerated
focus on plans for success and a reduction in the consideration of competing
tasks and other obstacles to successful completion. These cognitive processes

also played a mediating role in the enhanced optimistic bias found in the
incentive/motivated condition. This laboratory study serves as an ‘‘in vitro’’
examination of the processes underlying prediction; because the word
puzzle was a short, novel task, the optimistic bias was virtually absent in
the control condition, where students by and large directly reported the
highly salient past experiences from moments before. However, even in this
short and highly constrained laboratory setting, the presence of small finan-
cial incentives was sufficient to engage planning processes that overrode the
salient past experiences.
Similar effects of incentives for speed have been observed by other
researchers. Byram (1997) asked participants to estimate how long they
would take to perform an origami task in the lab. Participants offered a
monetary incentive for speedy completion of the task ($4 for finishing in the
top quartile, $2 for finishing in the next quartile, $1 for finishing in the third
quartile, and nothing for finishing in the bottom quartile) underestimated
their completion times, whereas participants without a monetary incentive
for speed did not exhibit the bias. Again, the incentive for speed increased
bias because it had a large impact on predicted completion times but did not
influence actual performance (for an economic analysis of the role of
incentives in the planning fallacy, see Brunnermeier et al., 2008).
Another strategy for exploring the role of motivational processes is to
solicit predictions from observers, who do not typically share the same
motivations as the individuals carrying out a task. Indeed, an early study
(Buehler et al., 1994, Study 5) indicated that observers are generally less
attentive than actors to actors’ reported plans and more attentive to potential
obstacles, the actors’ past experiences, and task deadlines—in other words,
observers are more likely to adopt the outside view and to construct a more
problem-focused inside view. Similarly, a more recent study found that
observers gave little weight to an actor’s motivation for early completion.
Actors who imagined themselves completing a school project predicted
they would finish it much earlier in a scenario where they were offered
incentive grades for early completion; however, knowing that the student
had been offered this incentive did not affect the predictions generated by
observer participants (Mulhern, 2006). Thus it appears that observers are
more skeptical than actors about whether incentives that operate on the
actor are enough to overcome past behavioral tendencies.
However, what would happen if observers were motivated to see an
actor achieve an early completion time? Conceivably, for such observers,
the same motivated reasoning processes that encourage optimism in actors
would operate so that they adopt an inside perspective dominated by plans
for success. We have tested this possibility by offering financial incentives to
observers that were contingent on a target (actor) individual’s timely task
completion (Buehler Griffin, 2009). In the initial study, actor participants
were given a take-home writing assignment to be completed within a

2-week deadline, predicted their completion time, and then carried out the
assignment. Observers were each yoked to a participant (actor) and received
the following information provided by the actor: demographic information
(sex, age, academic major), the instructions the actor received for complet-
ing the assignment, the actor’s report of typical assignment completion
times along with the self-predicted completion time for the target assign-
ment, and finally, the actor’s scenario-based description of how he or she
arrived at the prediction. We varied whether observers were offered a
reward that was contingent on the actor’s completion of the project. The
contingent observers were informed that they would receive a bonus
payment for participating in the study and that the size of the payment
would depend on how soon the target actor completed the assignment (50
cents for each day before deadline the assignment was completed). (Note
that these motivated observers were in no way compensated for the opti-
mism of their predictions nor were their predictions communicated to the
actors.) Neutral (control) observers were offered a standard bonus payment
that was not contingent on the actor’s performance. After reviewing the
information, observers were instructed to estimate as accurately as possible
when the target actor would finish the assignment and also rated the extent
to which they had based their predictions on the actor’s plans, past experi-
ences, and potential obstacles to early completion. They were also asked to
rate, specifically, the extent to which the actor’s past experiences were
relevant in predicting the completion time of the upcoming task.
The actor participants reported that they typically finished their assign-
ments about a day and a half before deadlines and, consistent with this report
of past behavior, finished the present assignment just prior to the 2-week
deadline (mean completion time was 12.1 days). However, in another
display of the planning fallacy, the actors predicted they would finish the
assignment in only 8.1 days, thereby exhibiting a substantial optimistic bias.
Observers in the contingent incentive condition, who were presumably
motivated to believe that the actor would finish as early as possible, also
tended to underestimate the completion time (mean prediction was 9.9
days) whereas neutral observers did not exhibit this bias (mean prediction
was 12.0 days). Furthermore, the thought ratings indicated that the incen-
tives affected observers’ predictions, in part, because they prompted the
observers to rely more heavily on the target individual’s future plans and to
place less weight on reports of past experiences and potential obstacles.
Participants’ direct ratings of the relevance of past experience also indicated
that incentives for early task completion led them to diminish the relevance
of past experience.
In a second study of contingent observers, we manipulated the extent to
which the contingent observers saw participants’ past experiences as rele-
vant using a variant of the recall-relevance manipulation used in our original
paper (Buehler et al., 1994). New observer participants were yoked to the

same actor participants as in the first study and were randomly assigned to
three conditions: neutral (control), contingent, and contingent plus past-
relevant scenario. In the new combined condition, observers were offered
the contingent incentive, but were also asked to indicate the completion
time suggested by the actor’s typical past experience and to generate a
scenario that could lead the person to finish at that time. Again, the ultimate
task was to estimate the target actor’s completion time as accurately as
possible. The study replicated the effect of incentive found previously,
and added the novel finding that forcing observers to confront the implica-
tions of an actor’s past behavior was enough to eliminate the incentive effect
on the planning process—and the corresponding degree of optimistic bias.
In a third observer motivation study, we constructed a set of ‘‘prediction
dossiers,’’ each about a hypothetical actor participant (Buehler Griffin,
2009). Four different actor profiles were constructed which differed in
terms of the individual’s typical task completion time relative to predictions
(early vs. late) as well as the current completion estimate (early vs. late). As in
the first two observer incentive studies, we manipulated whether or not
observer participants were offered a monetary incentive for each day before
deadline the apparently real task (an academic essay) was completed. Nota-
bly, the previously documented effects of contingent incentives on predi-
ction were not moderated by the profile information variables. In each
case, the observers who stood to gain from an early completion time
predicted the actor would finish earlier than did neutral observers. In
addition, thought-listings indicated that the effect of incentives was
mediated by the observers’ cognitive focus. Observers with an incentive
predicted earlier completion times because they focused more exclusively
on the actors’ optimistic plans, and relatively less on the actors’ previous
lateness or the potential obstacles they could encounter.
Taken together, these studies provide strong and consistent evidence
that directional motivation both increases the relative focus on the inside
versus the outside view, and accentuates the tendency to create a positively
biased inside view, as indexed by a narrow focus on plans for success and a
concomitant neglect of possible obstacles and interruptions. Notable is the
demonstration that even observers, who are usually less caught up in the
inside perspective than actors, can be made to ‘‘see’’ the world like actors
when they are motivated to hope for a short completion time.
5.3. Perspective(s) and the planning fallacy
Our work on motivation, described in the previous section, extends the
classic model of the planning fallacy by illustrating the important role of
motivation in altering the cognitive processes of prediction. The work also
touches upon two other important themes that are emphasized in our
extended model: first, that people who view the same future project from

different vantage points (e.g., an actor vs. observer perspective) can arrive at
very different conclusions, and second that even people who take an inside
view can vary markedly in the kinds of scenarios they construct (e.g., the
relative focus on plans vs. obstacles) and these variations within the inside
view can translate into different predictions. We take up each of these
themes further in this section by exploring variations in perspective or
vantage point that may influence the kinds of mental scenarios that people
construct.
5.3.1. First-person versus third-person imagery perspective
As people imagine themselves carrying out a future project they often
generate visual imagery (Atance O’Neill, 2001; Marks, 1999) and gener-
ally adopt one of two visual perspectives (Libby Eibach, 2002, 2009;
Pronin Ross, 2006). People who adopt a first-person perspective see the
project unfolding as if they were actually carrying it out. People who adopt
a third-person perspective see the events from an observer’s visual perspec-
tive; in their mental image, they see themselves as part of the field of view,
embedded in their surroundings. A defining attribute of a third-person
perspective is that individuals can ‘‘see themselves’’ acting, much like an
onlooker would. At a metaphorical level at least, the two visual perspectives
appear to map onto the inside and outside approaches to project prediction.
Furthermore, research on imagery perspective in a variety of domains
suggests two reasons why third-person imagery may interfere with the
cognitive and motivational processes that typically result in optimistic bias.
First, imagery perspective may affect people’s cognitive approach to predic-
tion. Adopting a third-person perspective elicits cognitive processes that are
more similar to those found in neutral observers than in actors (Frank
Gilovich, 1989; Pronin Ross, 2006; Robinson Swanson, 1993).
Robinson and Swanson (1993) proposed that the first- and third-person
perspectives activate cognitive processes typically associated with the
corresponding roles of actor and observer. For example, whereas observers
typically try to understand why a person is acting in a particular way, actors
are actively engaged in goal pursuit and thus focus on plans for accomplish-
ing their goals. Along similar lines, research on attribution has found that
individuals prompted to recall or imagine an event from a third-person
perspective make attributions that more closely resemble those typically
made by observers than those made by actors (Frank Gilovich, 1989;
Pronin Ross, 2006). Thus, people who imagine an upcoming task from
the third-person perspective may be less inclined to focus on plans and more
inclined to consider potential obstacles, and as a result would generate more
realistic predictions.
Second, imagery perspective may alter the salience and intensity of
people’s emotional engagement with a target outcome. People tend to
experience stronger affective involvement with an event when they

generate first-person, rather than third-person, imagery (Kross et al., 2005;
Lorenz Neisser, 1985; McIsaac Eich, 2002; Pronin Ross, 2006;
Robinson Swanson, 1993). For example, in first-person memories,
people report rich accounts of the affective reactions, physical sensations,
and psychological states they experienced during the event (McIsaac
Eich, 2002; Robinson Swanson, 1993). People who adopt a third-person
perspective remain relatively dispassionate, objective, and removed. In the
realm of planning and prediction, then, task-relevant feelings and motives
(such as the motivation to finish a task at the desired time) may be less vivid
and cognitively available to people who imagine the task from the third-
person perspective, and hence this perspective may be relatively immune
from the additional bias associated with motivated inference.
We recently conducted several experiments to explore the effects of
people’s imagery perspective as they envision an upcoming task (Buehler
et al., 2010b). The first two studies tested whether differences in imagery
perspective are indeed linked to the degree of bias in task completion
predictions. Participants first nominated a project they intended to complete
in the coming weeks, and were then instructed to visualize the project
unfolding: ‘‘Try to picture in your mind how the project is likely to unfold,
including details such as when, where, and how it will be done. For
example, you could picture the steps you will take to complete the project.
You could think of potential problems, interruptions, or distractions that
may arise and how these would affect you. Make an effort to really ‘‘see’’
how the project is likely to unfold.’’ In the initial study, participants then
reported whether they had adopted primarily a first-person perspective (i.e.,
seeing the events through their own eyes, just as they would when they
were actually occurring) or a third-person perspective (i.e., seeing them-
selves and their surroundings, just as an onlooker would). Approximately,
two-thirds of participants reported adopting first-person imagery and one-
third reported third-person imagery. In the next study, imagery perspective
was randomly assigned: participants were instructed to adopt either the first-
person or third-person perspective as they visualized the target project.
Immediately after the visualization procedure, participants predicted when
the project would most likely be finished and later reported their actual
completion times in an e-mail survey.
Both studies yielded a very similar pattern of results. Overall, predicted
completion times were significantly shorter than actual completion times,
indicating that, as usual, participants tended to underestimate how long their
projects would take to complete. However, the degree of optimistic bias in
prediction was related to whether participants had generated third-person
or first-person imagery. Participants predicted they would take significantly
longer to finish their tasks when they generated third-person rather than
first-person imagery. Actual completion times did not differ across condi-
tions, however, and thus the bias was reduced by third-person imagery.

Indeed, there was no optimistic prediction bias in the participants instructed
to adopt a third-person perspective.
The next study again manipulated imagery perspective and also intro-
duced measures to assess the focus of participants’ thoughts (plans vs.
obstacles) as well as their motivation to complete the task promptly. The
study also introduced a control condition (in which participants were not
assigned to a specified perspective) to locate the effects of visual perspective
more precisely. As expected, participants predicted they would take signifi-
cantly longer to finish the project in the third-person than in the first-person
condition. Predictions in the control condition fell in between, and did not
differ significantly from those in either manipulated condition. To examine
the role of cognitive focus, coders rated the extent to which participants
focused on their plans for successful task completion and, separately, the
extent to which they focused on potential obstacles they might encounter.
Comparisons across the conditions indicated that, as expected, participants
focused less on future plans, and more on potential obstacles, in the third-
person than in either the first-person or control condition. Furthermore,
mediational analyses indicated that the effect of perspective (third-person vs.
first-person) on prediction was mediated by the focus of participants’
thoughts as they generated predictions. Third-person imagery reduced the
tendency to focus narrowly on plans and increased the likelihood of con-
sidering obstacles, which, in turn, yielded longer completion predictions
(see Fig. 1.8).
In addition to assessing cognitive processes, we also explored the guiding
role of motivation. Two items that were included to assess participants’
motivation to complete the task promptly indicated that, as hypothesized,
participants experienced lower levels of motivation in the third-person con-
dition than in either the first-person or the control condition. Apparently,
–0.43*
(–0.35*) –0.15
Third-person
perspective
Predicted days
Focus on plans
(vs. obstacles)
–0.48*
Figure 1.8 Focus on plans (vs. obstacles) as a mediator of the effects of visual
perspective (third-person vs. first-person) on completion time predictions. The plan
focus index was based on ratings of focus on plans minus ratings of focus on obstacles.
Sobel z ¼ 2.98, *p 0.01 (Buehler et al., 2010b, Study 3).

third-person imagery served to dampen participants’ levels of motivationat the
time of prediction. Additionally, third-person imagery appeared to reduce
people’s tendency to base predictions on their feelings of motivation. Indivi-
duals who were feeling more motivated to complete the task quickly predicted
shorter completion times within the first-person condition (r ¼ .46,
p 0.01) and control condition (r ¼ .38, p 0.05), but not within the
third-person condition (r ¼ 0.10, ns). Similarly, regression analyses revealed
a significant interaction between motivation and perspective indicating that
motivation was a stronger determinant of predicted completion times in the
first-person than in the third-person condition (see Fig. 1.9).
Our final study in this series provided a fully experimental test of the
motivation by perspective interaction—that is, we experimentally varied
both the level of motivation to finish early as well as imagery perspective.
Participants imagined a project completion scenario, in which their profes-
sor had just given them a 15-page research essay due in 30 days. The
scenario included a motivation manipulation: half of the participants were
informed that, to encourage students to hand in the assignment promptly,
the professor was offering one bonus mark for each day before the deadline
the essay was submitted (high motivation condition), whereas the other
8
9
10
11
12
13
14
Low High
Motivation
Predicted
days
First-person Control
Third-person
Figure 1.9 Predicted completiontime for self-selected project as a functionofmotivation
to finish task and imagery perspective (Buehler et al., 2010b, Study 3).

participants were not informed of this incentive (low motivation condi-
tion). Participants were asked to visualize, from the assigned perspective
(first-person vs. third-person), the steps they would take to complete the
assignment, and then to predict when the project would most likely be
finished. The predicted completion times again revealed a perspective by
motivation interaction. Within the first-person condition, participants pre-
dicted they would finish much earlier in the high motivation condition than
in the low motivation condition. Within the third-person condition, this
effect of motivation was significantly attenuated. The results thus provide
further evidence that imagery perspective alters the role of motivation in
prediction.
Taken together, then, the studies of visual perspective supported three
guiding hypotheses about the effects of imagery perspective. First, people
predicted longer completion times—and thus were less prone to bias—
when they imagined an upcoming task from a third-person rather than a
first-person perspective. Second, the effect of imagery perspective was
mediated by people’s cognitive focus on plans versus obstacles at the time
of prediction. Third-person imagery reduced people’s inclination to focus
narrowly on optimistic plans, and increased their focus on potential obsta-
cles. Finally, imagery perspective altered the role of task-relevant motiva-
tion in prediction. People’s desire to finish an upcoming task promptly was
reduced, and weighted less heavily, when they adopted a third-person
rather than a first-person perspective. In essence, then, third-person imagery
elicited predictions and underlying psychological processes that mimic those
found previously in neutral observers (Buehler et al., 1994; Newby-Clark
et al., 2000).
5.3.2. Temporal perspective
In everyday life, people must generate task completion predictions at
various points in time; for example, a homeowner might try to predict
the time required for a renovation project months in advance (e.g.,
when booking the contractors) or when the work is about to begin (e.g.,
when arranging temporary lodging). Academic conference organizers often
deliberately invite participants far in advance when the prospect of giving a
talk seems a desirable and inviting chance to share knowledge and renew old
friendships. More generally, perceived temporal distance influences the
level of optimism for judgments and predictions in many other domains
(Liberman et al., 2007). Thus, we have begun to examine whether, and
how, the temporal distance to a future project influences people’s predic-
tions of when it will be finished.
According to temporal construal theory (TCT) (Liberman Trope,
1998; Trope Liberman, 2003), decreasing temporal distance enhances the
tendency for forecasters to focus on concrete, specific representations of an
event, whereas increasing temporal distance enhances the tendency for

forecasters to focus on high-level, abstract representations. For example, the
same future project (e.g., completing a school assignment) in the far future
will be construed at a high level of abstraction (e.g., pursuing educational
goals) but in the immediate future will be construed at a low-level or
concrete-level of abstraction (e.g., typing words on paper). High-level
construals contain information about the central and abstract features of
the event, but are schematic and decontextualized and hence do not contain
incidental or peripheral details. In contrast, low-level construals contain
more concrete, contextualized representations of the specific case at hand;
they are rich in detail, including information about incidental or peripheral
features of the event (Liberman et al., 2002).
In the realm of planning and prediction, people’s thoughts about both the
specific plans they hope to undertake and the specific obstacles they may
encounter (e.g., problems with the task, competing demands for their time,
changes in motivation, lack of skills or ability) may be heightened by
temporal proximity. Studies on predictions in other domains have found
that people become less optimistic and less confident in their predictions
about an event as it draws near (e.g., Eyal et al., 2004; Gilovich et al., 1993;
Savitsky et al., 1998); a plausible account for several of the findings is that
potential obstacles become more salient as an event draws near. For exam-
ple, Gilovich et al. (1993) found that students estimating their performance
on several short experimental tasks (e.g., recalling nonsense syllables) were
less confident and listed more possible reasons for failure when told they
would complete the tasks immediately rather than later in the semester.
Along similar lines, people appear to be more sensitive to competing
demands on their time in the near future than in the distant future
(Liberman Trope, 1998; Zauberman Lynch, 2005). These findings
imply that forecasters may focus more on obstacles to task completion (e.g.,
problems, interruptions, competing demands) when thinking about a task
that is closer in time.
Although temporal proximity heightens people’s focus on potential
obstacles, we postulate that it also increases the tendency to focus on
specific, concrete plans for carrying out a task at a desired time. Plans are
concrete in that they spell out the specific steps that will be taken to carry
out the upcoming task; they require that people move beyond abstract,
decontextualized representation of an event to consider features unique to
the case at hand. According to this logic, temporal closeness should prime or
make available concrete content of two different types, each with directly
opposite implications for prediction. In the context of a specific prediction,
situational factors may determine which of the two types of content is most
prevalent, and thus determine the size and direction of temporal proximity
effects on prediction. In circumstances where people are aware of or
concerned about obstacles they might encounter, temporal proximity may
accentuate these concerns, thereby resulting in less optimistic predictions.

In circumstances where people are already inclined to focus on developing
plans for success, with little concern about feasibility, temporal proximity
will strengthen the plan-based approach to prediction.
Our initial studies in this area (Peetz et al., 2010) find support for the
general pattern expected following TCT: people generate more optimistic
predictions when tasks are further in the future. For example, in an experi-
mental study, individuals predicted when they would finish the gift shop-
ping they planned to do for the upcoming Christmas holiday. All
participants came into the lab in October and provided their shopping list.
However, participants were then randomly assigned to make their predic-
tions immediately (for predictions made 3 months in advance of Christmas
day) or in a subsequent questionnaire session 6 weeks later (for predictions
made only 3 weeks in advance of Christmas day). Consistent with the classic
planning fallacy pattern, participants predicted they would finish their
nominated shopping earlier than they had typically completed their shop-
ping in previous years, but did not finish any earlier than in the past. Most
relevant to our present questions, participants in the temporally distant
condition predicted that they would finish their holiday shopping signifi-
cantly earlier than those asked closer to the holiday; because the two groups
did not differ in when they actually completed their shopping, the distant
predictions were more optimistically biased than the close predictions. Of
course this pattern could stem in part from differences in the time available
to carry out the shopping; however, in studies described below we have
replicated the effect of temporal proximity to a future task while controlling
for the time available to carry it out.
Thought focus measures shed further light on cognitive processes
underlying the effect of temporal proximity on participants’ shopping
predictions. Participants’ open-ended thought-lists were coded for refer-
ences to concrete obstacles (e.g., I figure with being in school I won’t have much
time to plan or go out and buy any gifts. So I gave myself some time once I have
completed all my exams to go out and buy them.) and to concrete plans (e.g., I
will go to a couple different malls in the area to look around). Perhaps in
recognition of the challenges of the holiday season, thought-listings
revealed a moderate focus on potential obstacles—indeed, there were
slightly more references to obstacles than to concrete plans. Although this
pattern may seem somewhat surprising, given that many previous studies
have found little focus on potential obstacles, the effects of this focus on
obstacles were clearly consistent with our theorizing. People mentioned
potential obstacles more frequently in the close condition than in the
distant condition, and a greater focus on obstacles, in turn, led people to
predict later completion (see Fig. 1.10). Thus, the impact of temporal
proximity on prediction was mediated, at least partially, by the degree of
focus on potential obstacles.

As developed above, the logic of temporal construal applied to our plan
focus construct implies that temporal proximity should sometimes yield
opposite effects. In particular, in those circumstances where people are
already inclined to focus more on optimistic plans, with little concern for
potential obstacles, temporal proximity should strengthen people’s tendency
to generate optimistic predictions. In order to test this reasoning, we tried to
create a context where there would be minimal concern about obstacles at
the time of prediction. We reasoned that for hypothetical tasks the obstacles
that tend to delay task completion (interruptions, problems, competing
demands) would not be salient (Armor Sackett, 2006) and instead people
would be unconstrained in constructing an optimistic scenario that depicts
the task unfolding as planned. To test this reasoning, in one study we varied
whether participants generated predictions for a task described as hypothet-
ical or real. Participants were asked to imagine completing (hypothetical
task) or to actually complete (real task) a take-home writing task that they
would receive either the next day (close condition) or not until the next
term (distant condition). In each case, the deadline was 2 weeks after
receiving the task, thereby controlling for the time interval available to
complete the task. Consistent with our theorizing, temporal distance had
directly opposite effects on prediction depending on whether the task was
hypothetical or real. For the hypothetical task, participants generated more
optimistic predictions when the task was closer in time, and this effect was
mediated by an increased focus on a concrete, plan-based scenario. For the
real task, in contrast, participants generated less optimistic predictions when
the task was closer in time, and this effect was mediated by an increased
focus on potential obstacles (see Fig. 1.11). It appears that temporal prox-
imity can increase the planner’s focus on potential obstacles to task comple-
tion or future plans for success, but has a more pronounced impact on
whichever type of cognition is focal as participants generate their
predictions.
A final experiment manipulated more directly individuals’ inclination to
focus on plans versus obstacles. Participants were again presented with the
0.10
0.09
−0.30*
0.31*
Predicted days
before deadline
Temporal
proximity
Focus on
obstacles
Focus on
plans
Figure 1.10 Focus on plans and obstacles as mediators of the effects of temporal
distance on completion time predictions for holiday shopping, *p 0.05 (Peetz et al.,
2010, Study 2).

hypothetical version of the writing task and imagined they would receive it
either the next day or 2 months hence. They were instructed either to focus
either on a step-by-step plan for completing the task (plan focus condition)
or possible problems and interruptions (obstacle focus condition) as they
read the detailed task instructions. According to our theorizing, predictors
should be more responsive to the manipulation of these concrete task
cognitions when they are contemplating a close task (and thus engaged in
concrete thinking) rather than a distant task (and thus engaged in more
abstract thinking). A significant interaction effect supported this reasoning.
For a task starting tomorrow, individuals predicted earlier completion times
when they were induced to focus on plans rather than obstacles, whereas for
a task starting in 2 months the concrete thought focus manipulation did not
affect predicted completion times.
Taken together, these findings supplement previous research and theory
concerning the role of temporal distance (Gilovich et al., 1993; Shepperd
et al., 1996) and hypotheticality (Armor Sackett, 2006) in other kinds of
performance predictions. Previous research demonstrates that people are
less likely to make optimistic predictions about their performance on an
imminent test than on a distant one. Our findings indicate that, depending
on the context, temporal proximity can elicit concrete thoughts about
either plans or obstacles. It may be that the effects of temporal distance
observed in previous studies are accounted for, in part, by contexts
(e.g., considering an imminent exam) that elicit a focus on obstacles over
Real writing task
Hypothetical writing task
0.01
0.05
0.33*
Predicted days
before deadline
Temporal
proximity
Focus on
obstacles
Focus on
plans
–0.32*
0.50*
0.29*
0.07
–0.05
Predicted days
before deadline
Temporal
proximity
Focus on
obstacles
Focus on
plans
Figure 1.11 Focus on plans and obstacles as mediators of the effects of temporal
distance on completion time predictions, *p 0.05 (Peetz et al., 2010, Study 4).

plans—one can no longer plan to study long hours if time has run out.
Our account also suggests the novel prediction that in contexts where step-
by-step planning is salient and feasible, temporal proximity can increase
optimism.
The present research also contributes to the expansive literature explor-
ing implications of TCT for judgment (for reviews, see Ledgerwood et al.,
2010; Liberman et al., 2007; Trope Liberman, 2003). Our findings offer
further evidence that increases in the temporal proximity of an event
heighten the role of concrete information in judgments about the event.
Moreover, our findings highlight the importance of distinguishing between
the types of concrete thought that may be enhanced by temporal proximity,
as they can have different, even opposing, effects on subsequent judgments.
Increased attention to different types of concrete cognitions and their
distinctive consequences may yield valuable insights into the effects of
distance in other judgment domains as well.
A quite different approach to studying temporal perspective was taken
by Sanna et al. (2005), who varied perceived distance by focusing forecasters
on how a standard period of time was either short or long. Using class
projects with a 3-month deadline as the target task, a negative temporal
frame was created by reminding some groups of forecasters that ‘‘given the
deadline. . ..you only have 12 weeks remaining’’ and a positive temporal
frame was created by reminding other groups of forecasters that ‘‘beginning
today. . .you still have 12 weeks remaining.’’ Consistent with the classic
inside–outside analysis of the planning fallacy, predictions made under a
positive temporal frame were essentially identical to those made in a control
condition, whereas the negative temporal frame led forecasters—relative to
the other two conditions—to rate the deadline as perceptually closer, to rate
thoughts about success as more difficult to bring to mind, and to make later
and less biased forecasts for completion time. The same pattern was repli-
cated in a laboratory experiment using furniture assembly as the target task.
Sanna et al. (2005) also manipulated the number and valence of thoughts
that forecasters were required to list and found that optimistic predictions
were more marked when success thoughts were experienced as easy to
bring to mind (because participants were only asked to list a few) and
thoughts about problems were perceived as difficult to bring to mind
(because participants were asked to list many). Conversely, optimistic pre-
dictions were attenuated when success thoughts were difficult to bring to
mind and thoughts about problems were easy to bring to mind. This latter
demonstration indicates that under certain experimental conditions the acces-
sibility versus frequency of positive versus negative thoughts—essentially what
we have called positive future focus—can be dissociated (see Sanna
Schwarz, 2004, for a similar effect). Furthermore, under these conditions,
where ease of generation and amount of content are negatively correlated,
perceived ease was the driving force. How this translates into a naturalistic

setting, when ease and overall valence of thoughts are positively correlated,
is an interesting area for future research.
5.4. Social forces: Group processes accentuate plan focus
Consulting with and working with others is a necessary part of many
activities within organizations, including forecasting (Armstrong, 1985).
Even when project managers render a final prediction alone, they usually
interact with others to collect information, opinions, and advice (Heath
Gonzalez, 1995). Similarly, people often consult family members or friends
when predicting their personal futures (MacDonald Ross, 1999; Shrauger
et al., 1996). How does this affect the optimistic bias associated with the
planning fallacy? On one hand, the group discussion process brings with it
an opportunity to consult an outside-oriented perspective; on the other
hand, the group discussion may bring an even greater focus on success and
reduce or eliminate the influence of the outside perspective.
In one research stream, we have examined tasks performed collabora-
tively by groups, and predictions generated through group discussion
(Buehler et al., 2005). Participants predicted—individually and collabora-
tively—when they would complete upcoming group projects ranging from
brief laboratory tasks to extensive real-world projects, and their actual
completion times were measured. The research was guided by three specific
questions: Are completion forecasts for group tasks optimistically biased?
Do groups generate predictions for group tasks that are more or less biased
than those generated by individual group members? How does interaction
and discussion influence the degree of future-focused thought and hence
the degree of bias in predictions generated by—or least influenced by—
group process?
There is some reason to expect that optimistic biases found for individual
tasks may be reduced or eliminated for group-based tasks. An important
feature of truly interactive group tasks is that people are interdependent—
they must often wait for contributions of others before they can do their
own part. In some ways, then, participants’ viewpoints are similar to that of
an observer of other people’s activities, which might give rise to an outside
perspective and mitigate the planning-based optimistic bias. However,
membership in a collaborative group—even an ad-hoc group such as
those we study—also creates important differences in perspective relative
to neutral observers. Although a group member may be partially an observer
(for the other members’ contributions), he or she is also an active participant
(for his or her own contribution) and has a stake in the task as a whole. We
reasoned that the role as an active contributor and stakeholder to a project
would override any tendency to think as an outside observer—and thus we
hypothesized that people’s predictions would be optimistically biased for
group tasks.

If group predictions are optimistically biased, should we expect a smaller
or larger error than for individual predictions? Such comparisons are impor-
tant because a common justification for consulting with others is to increase
prediction accuracy. However, relevant research in other judgmental
domains has yielded inconsistent evidence about whether interaction with
others improves the quality of judgment (for a review, see Kerr Tindale,
2004). Although some studies have indicated that group discussion attenu-
ates judgmental biases (Wright Wells, 1985; Wright et al., 1990), several
other findings suggest that groups tend to accentuate rather than attenuate
judgmental biases found in individuals (Hinsz et al., 1997b; Tindale, 1993;
Tindale et al., 1996). Indeed, the ‘‘group accentuation effect’’ noted by
Tindale and colleagues affects two judgmental biases—overconfidence and
base-rate neglect—that are close cousins to the planning fallacy (Argote
et al., 1986, 1990; Heath Gonzalez, 1995; Tindale, 1993). On the basis of
this research, we hypothesized that group discussion will tend to heighten or
accentuate the optimistic bias in people’s task completion forecasts.
Our third question concerned the mechanism or process by which
collaboration may result in overly optimistic forecasts. Prior theoretical
work on group-mediated shifts in attitudinal judgment (Hinsz et al.,
1997b; Tindale, 1993; Vinokur Burnstein, 1974) suggests that group
discussion prompts individuals to focus selectively on information that
supports or accentuates the initial inclination of the individual group mem-
bers. In the present judgment domain, we have argued that people’s initial
inclination is to focus narrowly on thoughts of successful task completion.
We expected that group discussion would reinforce this tendency, prompt-
ing group members to focus even more exclusively on task-relevant infor-
mation (plans for task completion, qualities of the task, qualities of the group
members, qualities of the situation) which imply that the task will be
completed promptly, and consequently to neglect or underweight informa-
tion suggesting a task may take longer than desired.
Our first study (Buehler et al., 2005) provided an initial test of the
hypothesis that forecasted completion times for group-based projects tend
to be optimistically biased, using a real academic task. Undergraduate
business students predicted when they would complete each of the four
components of an extensive (9-week) required course project carried out in
small groups, and later reported when the work was finished. The study also
provided a preliminary comparison of individual and group-based predic-
tions using a simple pretest/posttest design. Individual predictions were
generated by each group member and later each group arrived at a single
group-based prediction through discussion. As predicted, participants
underestimated their completion times for each phase of the project, and
this optimistic bias appeared both for predictions generated by individual
group members, as well as predictions generated collaboratively. In addition,
for each task component, the group-based prediction was more optimistic

than the individual predictions. Notably, all four components were com-
pleted in the last 10 days before the project was due, on average, whereas
predictions based on group discussion indicated that the first component
would be complete about 6 weeks before the due date, more than a week
earlier than indicated by the initial individual predictions. Similar tasks were
studied by Sanna et al. (2005), who confirmed that the classic optimistic
prediction bias was robust for group projects but did not compare group and
individual predictions.
In two additional studies, we examined the consequences of group
discussion using more tightly controlled laboratory experiments. The target
tasks were a jigsaw puzzle task completed in the lab and a take-home
assignment that involved finding answers to factual questions. In each
case, to ensure that participants desired an early task completion, they
were provided with incentives for finishing quickly. In the puzzle task
study, participants were also provided with the average time taken in a
separate pilot study. Group members were randomly assigned to generate
task completion predictions under one of three discussion conditions:
consensus, consultation, and an individual (no discussion) condition. In
the consensus condition, groups of participants arrived at a single prediction
that represented the consensus of the group. In the consultation condition,
participants discussed their predictions with fellow group members before
rendering a final prediction privately. Although we did not anticipate
differences between these two conditions, we sought to determine whether
effects of group discussion occurred primarily due to consultation or due to
the requirement of negotiating a common group forecast. We also varied
whether group members made initial private predictions prior to the dis-
cussion manipulation. Participants in all conditions were asked to list the
thoughts and ideas they had considered or discussed to arrive at their
predictions.
In both studies, participants in all conditions underestimated how long
they would take to complete the tasks, but predictions based on group
discussion were significantly more optimistic (and hence more biased) than
predictions generated by individual group members, whether that compar-
ison was made over time within participants (comparing sequential individ-
ual and group-based predictions) or across conditions (comparing individual
and group discussion conditions). There were no differences between the
Consultation and Consensus conditions, indicating that it was the group
discussion, rather than the process of agreeing on a specific forecast, that was
the key factor. Notably, for all conditions, the actual completion times on
the puzzle task were very close to the normative time reported to partici-
pants before they made their forecasts (14 min and 10 s); individuals on
average incorrectly expected they could shave a full minute off that time,
and group discussion led to the removal of an additional minute.

The thought-listing measures allowed us to test whether the group
discussion effects were mediated by the focus of participants’ thoughts as
they arrived at their predictions. To this purpose, the verbal protocols were
coded into four main categories (future scenarios, past experiences, group
qualities, and task characteristics) and in each case the thoughts in these
categories could be either positive or negative (i.e., imply a relatively short
or long completion time, respectively). This coding scheme differs some-
what from those we have used in previous studies. It was developed to suit
group-based tasks and to allow a closer examination of the role of positive
versus negative thought content.
The findings suggested that group discussion effects were mediated by
the positivity of participants’ thoughts as they generated their final predic-
tions. Group discussion prompted an increased focus on several factors
promoting successful task completion, and this positive future focus
mediated the effects of group discussion on prediction. This mediation
effect was found for an overall index of the extent to which participants
focused on the positive versus negative forecasting content. However,
analyses of the individual coding categories also suggested that the pattern
of mediation was most pronounced and reliable (appearing across both
studies) for the category involving future scenarios. Group discussion
increased participants’ focus on positive future scenarios (i.e., scenarios in
which things go according to plan) and decreased their focus on negative
future scenarios (i.e., scenarios that included possible obstacles or problems),
and this pattern of thought focus resulted in more optimistic forecasts.
In sum, the studies supported our three guiding hypotheses about pre-
dictions in group contexts. We found a consistent optimistic bias in time
predictions for collaborative group projects (for further evidence of this bias,
see Sanna et al., 2005), we found that predictions based on group discussion
were more optimistic than individual predictions, and we found that the
effects of group discussion were mediated by individual group members’
thought focus at the time of prediction. These findings extend the previous
evidence of optimistic bias in important ways—to include target tasks
carried out by groups and to include group-based predictions—and also
have considerable applied value. Planners in organizational contexts often
develop their forecasts in team meetings and are called upon to justify their
predictions on the basis of detailed step-by-step plans. Presumably people
engage in consultation and discussion because they believe it will improve
the accuracy of their forecasts. According to the results of our studies,
however, group discussion can actually exacerbate the tendency toward
unrealistic predictions through an even greater tendency to ‘‘plan for
success.’’ These findings may be applicable to many collaborative work
ventures that are similar in form to the projects examined here—that is,
projects to be completed collaboratively by groups who are motivated to
finish early. Under such circumstances, forecasters may be well advised to

collect and aggregate individual forecasts, instead of engaging in group
discussion. It may also be beneficial to consult neutral observers, who appear
to be less prone to optimistic biases (e.g., Buehler et al., 1994), rather than
relying solely on predictions from active contributors.
Other social factors also appear to affect the optimistic time prediction
bias. One study tested the possibility that self-presentation concerns may
influence people’s task completion predictions (Pezzo et al., 2006b). Parti-
cipants generated more optimistic predictions when they reported them
verbally to an experimenter than when the experimenter was absent and
predictions were anonymous. Another recent program of research suggests
that people’s social roles may influence how they generate task completion
predictions. Weick and Guinote (2010) proposed that people who occupy
positions of power in the social hierarchy—and experience the sense of
power associated with that role—may be particularly prone to optimistic
bias. This hypothesis is based on previous evidence that power induces goal-
directed attention and a tendency to disregard information that lies outside
the focal goal (Guinote, 2007a,b). Applied to the planning fallacy, then,
powerful individuals should be particularly inclined to focus narrowly on
planning for success and to disregard potential obstacles. Consistent with
this reasoning, a series of experiments by Weick and Guinote indicated that
individuals were more likely to underestimate task completion times when
they were induced to experience feelings of power. Additional findings
supported the hypothesis that the effect of power on prediction was
mediated by attentional focus: Drawing participants’ attention to relevant
information outside the focal goal (i.e., relevant past completion times)
eliminated the effect of power on prediction.
5.5. The behavioral impact of plans and predictions
We have described a number of extensions of the planning fallacy
model, and we have demonstrated that each of these variables works
through the relative balance of positive and negative content of the inside
planning process; this balance, in turn, influences the optimism of predicted
completions and, hence, affects the degree of bias. So far, in developing this
model, we have explicitly assumed that the influences of plans on behavior
are small relative to their effects on prediction. Furthermore, we have
implicitly assumed that any effects of predictions on behavior are similarly
small.
However, in some circumstances, the prediction—and the planning
process that gave rise to it—may have wide ranging consequences on how
the predictor feels, thinks, and acts; in particular, an optimistic prediction
may have self-fulfilling consequences that may be quite adaptive for the
predictor. This is particularly true for the personal projects that we study,
which may languish uncompleted because of the forecaster’s self-control

problems and lack of concrete goals. This possibility is less relevant for
applied projects such as software engineering or large-scale public works,
where no amount of wishing, goal-setting, and motivation can make a
project less complex or expensive. (It has been argued that simply adding
personnel to a behind-schedule software development project may actually
lead to a slower completion cycle because of training and coordination costs
(Brooks, 1975, p. 25); Brook’s Law states that ‘‘Adding manpower to a late
software project makes it later.’’)
Within the domain of personal projects, people’s predictions may some-
times influence when and how they carry out the target tasks. Imagine, for
example, a student attempting to predict when she will finish an extensive
project due at the end of the month. Although painfully aware that she
typically finishes such projects just before the relevant deadline, she hopes
and intends to complete this one well ahead of schedule and, reflecting this
honest hope, estimates that she will finish a week before it is due. Could it
be that the very act of generating an optimistic forecast will help her to finish
the task, if not by the predicted time, at least earlier than she would have
otherwise?
Such questions are not easy to answer empirically. It may be tempting to
examine the association between people’s predicted and actual completion
times and, as noted previously, many past studies have found strong positive
correlations. These correlations are at least consistent with the idea that
predictions influence actual completion times: They indicate that indivi-
duals who predict to finish a task earlier (relative to other individuals in the
sample) do end up finishing relatively early. However, they do not provide
conclusive evidence that the earlier predictions actually cause earlier com-
pletion times. Such a conclusion requires an experimental manipulation of
predicted completion times to examine their impact on behavior.
In our research, we have begun to explore the questions of whether, and
when, task completion predictions have a behavioral impact (Buehler et al.,
2010a). These questions bear some resemblance to issues addressed in
research on goal-setting and task motivation, which concludes that specific,
challenging goals can enhance performance in many tasks (Locke Latham,
1990, 2002). However, generating predictions and setting goals are distin-
guishable acts: Whereas goals represent desired future outcomes that people
strive to attain, predictions represent beliefs about what will actually tran-
spire. Certainly people often set goals for project completion, but there are
also many circumstances in which they strive to predict as accurately as
possible when a project will actually be done. They may be called upon by
others—such as supervisors, coworkers, or spouses—to provide a realistic
estimate. They may privately seek an accurate estimate to guide their own
plans and decisions. In such circumstances, it would be valuable to under-
stand whether the forecasts people generate are likely to guide their actual
completion times.

Several theoretical approaches suggest that optimistic predictions could
facilitate early task completion times. An early task completion prediction
might help to initiate action because: it serves as a guiding goal or standard
(Locke Latham, 1990, 2002); it operates like other performance expec-
tancies to increase people’s confidence, effort, and persistence (Bandura,
1977; Carver Scheier, 1998; Cervone Peake, 1986); it induces consis-
tency motives (Festinger, 1957); it elicits concrete plans or scenarios that
specify when, where, and how the task can be completed promptly (Armor
Taylor, 2003; Gollwitzer, 1999; Taylor et al., 1998); and it heightens the
accessibility of the person’s attitudes and intentions toward the behavior
(Morwitz Fitzsimons, 2004; Sherman, 1980). However, despite these
potential mechanisms, research summarized here suggests that predictions
themselves often have little or no influence on completion times.
A challenge, then, is to reconcile two apparently contradictory findings:
first, that expressed predictions can affect subsequent behavior (e.g.,
Morwitz Fitzsimons, 2004; Sherman, 1980), and second, that manipula-
tions that indirectly affect the optimism of predicted task completion times
have typically failed to show measurable behavioral effects. We propose that
task completion predictions are more ‘‘translatable’’ (Koehler Poon,
2006) into the one-time action of starting a project than into the continuing
actions necessary to complete a project. This assumption has clear implica-
tions for identifying the type of tasks for which optimistic predictions will
have beneficial effects. In particular, tasks that can be completed in a single,
continuous session (e.g., reading a magazine article, writing a thank-you
note, changing a lightbulb), and thus are not usually prone to interruptions
and delays once they have been started, are the type of tasks where we
expect an effect of predictions on completion times. We term these ‘‘closed
tasks’’ because once initiated, these tasks are relatively insulated from outside
disruption. In contrast, ‘‘open tasks’’ require multiple steps to be completed
at different times or locations. These are the kinds of tasks that typically
require several work sessions to finish (e.g., reading a textbook, writing a
manuscript, completing a home renovation) and thus are relatively prone to
interruptions and delays even after they have been started; hence, for these
tasks, differential plans, intentions, and starting times may not influence the
ultimate completion times.
Consider, for example, a homeowner who wants to hang a newly
acquired painting in the hallway, a task that requires purchasing hooks,
fastening them to the wall, and hanging the artwork. There are only a few
easy steps (i.e., the task is quite simple) and the actual working time should
be less than an hour (i.e., the task is brief in duration), yet this is an open task
because the steps are spread out across multiple times and places, making it
prone to external interruptions. Even after starting the task, the individual
could be easily sidetracked, and days or weeks might pass before the painting
is on display.

Our reasoning, then, is that a task completion prediction will exert its
greatest impact on the early stages of task completion, but this impact will
diminish as delaying factors are encountered. Consequently, the effects of
predictions on task completion times will differ for closed and open tasks.
For a closed task, an early start time should be sufficient to influence final
completion times because there is little opportunity for delay once the task
has been started. For an open task, however, an early start is much less likely
to translate into an early completion time because there is great potential for
delaying factors to intervene between starting and finishing the task.
As one test of our theoretical framework, we performed a quantitative
review of studies that have examined the impact of various factors on
predicted and actual completion times (Buehler et al., 2010a). Recall that
this research has often found that factors affecting prediction do not have a
corresponding effect on behavior. But according to our reasoning, the
extent to which this pattern is obtained should depend on whether the
tasks are open or closed. Thus, we identified studies in which researchers
manipulated a factor to influence predicted completion times and reported
means for both predicted and actual completion times. Two research
assistants independently coded the target tasks as being either relatively
closed (e.g., a 1-h computer tutorial, a short writing assignment) or rela-
tively open (e.g., major school projects, income tax returns). We then
examined the effect of each manipulation (i.e., the mean difference across
conditions) on predicted and actual completion times and created a behav-
ioral effect index by dividing the effect on behavior (i.e., the mean differ-
ence between conditions in actual completion time) by the effect on
prediction (i.e., the mean difference between conditions in predicted com-
pletion time). This index represents the extent to which an effect on
prediction was accompanied by a corresponding effect on behavior,
where a score of 0 indicates a complete lack of a behavioral effect, and a
score of 1 indicates that the effect of the manipulation on behavior was equal
to its effect on prediction, and a score greater than 1 indicates that the effect
of the manipulation on behavior was greater than its effect on prediction.
The behavioral effect index scores were significantly higher for closed tasks
than for open tasks (Ms ¼ 0.76 vs. 0.31), t(12) ¼ 2.60, p ¼ 0.02, indicating
that differences in prediction were more likely to carry through to behavior
for closed tasks than for open tasks.
To supplement the findings of this systematic review, we also supervised
a study designed to assess the impact of prediction on completion behavior
(Deslauriers, 2002). In this study, some participants were instructed to think
about (and describe) the specific details of when, where, and how they
would complete the target assignment (plan focus condition), and others
were not given this instruction (control condition). The extent to which the
target task was open or closed was also manipulated, providing a test of
whether differences in prediction induced by a plan focus would carry

through to affect the actual completion times of open and closed tasks. All
student participants were given a take-home assignment that consisted of
writing three short essays (500 words each) and submitting the essays by
e-mail as soon as they were completed. To manipulate the extent to which
this writing assignment constituted a closed or open task, participants were
told either that they could complete all three of their essays in one sitting, as
long as they submitted each one individually (closed task condition) or that
they were required to let 1 day elapse between writing and submitting one
essay and starting to compose the next one (open task condition). Thus,
although the actual work required for the assignment was identical in each
condition, the latter version of the assignment constituted an open task
because the steps of the assignment needed to be completed across multiple
work sessions and, consequently, would be relatively prone to interruptions
and delays caused by factors external to the task itself.
An examination of predicted and actual completion times for the assign-
ment revealed a pattern of effects that was consistent with our theorizing.
Although all participants tended to underestimate their task completion
times, the degree of bias differed markedly across conditions. For the
open version of the task, focusing on plans increased the optimistic bias:
Participants predicted to finish the task in fewer days when they were
induced to focus on plans than when they were not (Ms ¼ 7.4 vs. 8.9
days) but their actual completion times did not differ (Ms ¼ 10.4 vs. 10.3
days), and thus underestimation was greater in the plan focus than in the
control condition (Ms ¼ 3.0 vs. 1.4 days). In other words, the effect of plan
focus instructions did not carry through from prediction to affect actual
completion times. However, a very different pattern was seen for partici-
pants assigned to the closed version of the task. Participants again predicted
to finish in less time in the plan focus condition than in the control
condition (Ms ¼ 5.8 vs. 7.7 days); however, the actual completion times
were also shorter in the plan focus than in the control condition (Ms ¼ 7.9
vs. 10.2 days), and thus the bias was similar across the two conditions
(Ms ¼ 2.1 vs. 2.5 days).
Although the pattern of effects observed in this study is generally
consistent with our reasoning about the moderating role of open versus
closed tasks, note that the studies so far did not manipulate predictions
directly, and thus it is plausible that the plan manipulation influenced both
predicted and actual completion times, without a causal role of prediction
on behavior. To assign a causal role to prediction, it is desirable to manipu-
late predictions as directly as possible. Thus in a series of experiments
(Buehler et al., 2010a), we used anchoring procedures to manipulate parti-
cipants’ task completion predictions more directly and then examined their
actual completion times on a variety of open and closed tasks. The specific
anchoring manipulations varied across studies, from drawing a card to
sliding a pointer along a numbered time line, but each one provided

participants with an ostensibly ‘‘random’’ starting point that was either early
or late. Participants were then asked to adjust from this starting point to
arrive at their predictions for a particular project. In each study, the anchor-
ing manipulation yielded a strong effect on predictions, allowing us to
investigate whether the completion times would follow predictions. The
type of task (open vs. closed) was varied across studies as well as experimen-
tally manipulated within a study. We tested three central hypotheses derived
from our theorizing: first, that a manipulation of task completion predic-
tions can influence actual completion times; second, that manipulating
predictions will result in stronger behavioral effects on completion times
for closed tasks than for open tasks; third, that even for open tasks manip-
ulating predictions will result in behavioral effects on start times.
The results were consistent with each of these hypotheses. First, at least
for tasks that were closed (e.g., a computer tutorial assignment, a simple
writing assignment), we found that manipulated differences in prediction
carried on to affect actual completion times. For example, for the computer
tutorial assignment (which could be completed easily in one session),
participants exposed to the early anchor forecasted that they would finish
5 days earlier than those exposed to the late anchor, and actually did finish
about 5 days earlier (although both groups showed a substantial optimistic
bias). Second, we found that predictions had a greater impact on the actual
completion times of closed tasks than of open tasks: indeed, for each of the
open tasks that we examined (e.g., major school projects, income tax
returns), completion times were not affected significantly by the prediction
manipulation. Third, even though predictions did not influence completion
times for open tasks, the predictions did influence when these tasks were
started. On the tax return task, for example (see Fig. 1.12), participants
exposed to the early anchor predicted that they would send in their returns
almost 20 days earlier than those exposed to the late anchor, and actually
started the task some 18 days earlier, but eventually completed the task only
5 days earlier (a nonsignificant difference between conditions). These find-
ings are consistent with the theory that completion predictions have their
greatest impact on the beginning phases of a project, particularly the initia-
tion time, and that this effect diminishes over the course of an extensive,
multistage project.
The results of these studies not only support our theoretical framework
but also have practical implications. They begin to address the complex
questions of whether, or when, it may be adaptive to predict an early task
completion time. People hoping to finish a task early may be well advised to
predict an early completion time if the task is relatively closed, because the
predictions appear to facilitate early completion times without increasing
proneness to underestimation (and any associated costs). When the task is
relatively open, people may be advised to make more conservative forecasts,
especially if there are serious costs associated with underestimation, because

the predictions appear to have little benefit for actual completion times but
have clear potential to exacerbate bias. When an earlier starting time has
value in itself, then optimistic forecasts may be valuable regardless of the
nature of the task. The findings also provide guidance for inducing desired
task completion times in others, at least when tasks are relatively closed.
Techniques used in the early anchor condition of our studies could be
adapted for use in a range of applied settings where practitioners (e.g.,
teachers, managers, coworkers) wish to encourage others to complete
tasks well before a final deadline. A potential benefit of the early anchor,
relative to simply imposing a goal on others, is that it may be less likely to
jeopardize the individual’s commitment to finish at the desired time (Hinsz
et al., 1997a).
We believe that to fully understand the pattern of moderation, it is
necessary to consider the role of deadlines in guiding task completion
behavior. Note that each of our target tasks had a firm deadline for comple-
tion, and previous research has shown that such deadlines exert a powerful
impact on when people actually finish their tasks—even when people hope
to finish well in advance of a deadline, their actual time ends up being driven
largely by deadlines (Ariely Wertenbroch, 2002; Buehler et al., 1994).
We suspect that for open tasks, the salience of an early prediction (and
perhaps the task itself) may fade until a looming deadline pulls it back into
focus. That is, people’s final task completion times may be a function of two
0
5
10
15
20
25
30
35
Prediction Start Completion
Days
before
deadline
Early anchor Late anchor
Figure 1.12 Predicted completion time, actual start time, and actual completion time
for income tax forms as a function of anchor condition (Buehler et al., 2010a, Study 4).

psychological forces—people’s plans to finish early and their deadlines.
Optimistic plans to finish early may get tasks started, but then the effects
of external factors take over and it is the force of deadlines that controls
actual completion times.
5.6. Implications for debiasing
Although we have focused on exploring the scope and origins of the
planning fallacy, the extended model also has important practical implica-
tions, particularly for those seeking to ‘‘debias’’ task completion predictions
(for previous reviews of debiasing, see Buehler et al., 2002; Roy et al.,
2005). Below we briefly highlight implications of the model, and the
empirical research supporting it, for reducing people’s tendency to repeat-
edly underestimate task completion times.
A central implication of both the classic and extended inside–outside
model is that people can reduce optimistic bias by adopting an outside view,
wherein they use past distributions as the basis for prediction rather than
relying on goal-based plans and scenarios for the future. A direct application
is seen in Reference Class Forecasting, an approach to forecasting described
by Kahneman and Tversky (1979), named by Lovallo and Kahneman
(2003), and tested in the realm of major infrastructure projects (Flyvbjerg,
2008; Flyvbjerg et al., 2009). Reference class forecasting sidesteps the
problems associated with an inside view, as well as people’s disinclination
to base predictions on past experience, by requiring forecasters to base their
predictions on the outcomes of a distribution of comparable projects. The
empirical tests of this approach support its effectiveness in reducing time and
cost overruns in large-scale construction projects. Similarly, on a smaller
scale, our research on individual projects showed that an intervention
designed to induce individuals to base predictions on past experience (by
highlighting the relevance of their typical completion times to the target
project) yielded unbiased predictions (Buehler et al., 1994). Of course
forecasts based on past experience may still be overly optimistic to the
extent that people believe previous tasks took less time than they actually
did (Roy et al., 2005, 2008). However, given that individuals often hold
relatively pessimistic views of their own previous completion times, we
believe interventions that induce people to base predictions on past experi-
ence (such as reference class forecasting) could often help to debias
prediction.
What about those instances where forecasters do adopt an inside
approach to prediction? According to our extended model, their predic-
tions will still vary greatly depending on factors that determine the extent to
which they focus on goal-based plans or on obstacles, and thus we believe
each factor identified in the model could be applied strategically to reduce
or eliminate prediction bias. The findings of actor–observer differences

imply that prediction bias could be reduced by consulting neutral outside
observers. Such observers should be less likely than the project stakeholders
to focus narrowly on goal-based plans and more likely to account for
potential obstacles to task completion (Buehler et al., 1994; Mulhern,
2006; Newby-Clark et al., 2000). Our research on imagery perspective
implies that similar benefits might accrue from adopting a third-person
perspective when imagining a future task (Buehler et al., 2010b). The
guided imagery procedure used in these studies could be readily adapted
for use as an intervention in applied contexts, with a number of advantages
over other debiasing strategies. For example, unlike strategies that rely on
previous experience, it does not require that individuals have accurate
knowledge about relevant previous tasks, and unlike strategies that alter
the kinds of information people consider, the imagery perspective interven-
tion is not content-specific, and thus could be applied generically to a wide
range of target tasks and contexts.
Our research on temporal perspective tells a somewhat complex story
but, again, highlights the importance of attending to potential obstacles
while generating predictions (Peetz et al., 2010). In one study, for instance,
a manipulation that induced people to focus on obstacles, rather than goal-
based plans, led them to predict later completion times, particularly for
projects in the distant future. This suggests that interventions that focus
people’s attention, directly and explicitly, on potential obstacles may help
them to generate realistic predictions for future tasks. Our extended model
also highlights the role of motivational and social processes in the planning
fallacy, suggesting that people may be able to generate more realistic fore-
casts in contexts that minimize their motivation to finish early, or at least the
salience of such motives, at the time they are generating predictions, and
also in contexts that reduce social pressures that might encourage optimistic
predictions (Buehler et al., 2005; Pezzo et al., 2006b).
Another debiasing strategy that relies on the content of the inside-
focused scenario involves the ‘‘unpacking’’ or ‘‘segmentation’’ of a target
task into smaller units (e.g., Connolly Dean, 1997; Forsyth Burt, 2008;
Kruger Evans, 2004). As noted previously, unpacking a whole project
into subcomponents is likely to result in less optimistic predictions when the
subtasks serve as reminders of steps that need to be completed, but would
have otherwise been neglected in a scenario-based project plan (Kruger
Evans, 2004). In some respects, then, the recognition of additional steps to
task completion is very similar to the recognition of obstacles that may delay
task completion, and consistent with evidence that focusing on such cogni-
tions can attenuate the planning fallacy.
Finally, as our research on the behavioral impact of predictions illu-
strated, the degree of bias in prediction is not determined solely by processes
occurring at prediction, but also depends greatly on what transpires after a
prediction has been generated. Thus one way to increase the accuracy of

prediction is to bring people’s behavior in line with the predictions they
have generated. For example, given that people’s focus on their goals and
plans (salient at the time of prediction) may diminish across time, one
remedy would be to enhance the salience of goal-directed plans throughout
the entire project period. In a relevant test of this idea (Taylor et al., 1998),
students were trained to imagine themselves performing each of the steps
needed to complete a school-related project by their predicted completion
time. Participants rehearsed these simulations each day of the week during
which they carried out the project. The exercise increased the proportion of
students who finished by the predicted time from 14% (in a control
condition) to 41%. Along similar lines, Koole and Van’t Spijker (2000)
found that students were more likely to finish a brief writing task by the
expected time if they were instructed to form ‘‘implementation intentions’’
(Gollwitzer, 1999; Gollwitzer Brandstätter, 1997) that specified the exact
time and place they would do the task. These findings suggest that one way
to reduce prediction bias is to target the behaviors that occur after predic-
tion. Our recent research suggests that such behavior-based approaches may
vary in effectiveness depending on the nature of the task, and particularly
the extent to which it is prone to unexpected interruptions, problems, and
delays (Buehler et al., 2010a).
6. Concluding Perspectives
The planning fallacy is a phenomenon rooted in daily life and daily
experience, rich enough to keep our research group busy for the best part of
two decades. Although this chapter explored the planning fallacy in the
domain of task completion predictions, we have also documented similar
patterns and psychological processes in people’s tendency to overpredict the
longevity of their romantic relationships (Buehler et al., 1995; see also
MacDonald Ross, 1999) and to overestimate the intensity of their
emotional reactions to upcoming events (Buehler McFarland, 2001). In
each case, we find that people’s natural inclination to take the inside
approach to prediction, and to overlook or dismiss the outside approach,
leaves them prone to bias. A recent program of research extends our work
on the planning fallacy into the realm of personal financial behavior (Peetz
Buehler, 2009). In this domain we again find that a similar phenome-
non—a ‘‘budget fallacy’’—exists for personal spending predictions. Indivi-
duals tend to underestimate their future spending, predicting that they will
spend less money in the upcoming week than they actually spend, and also
predicting that they will spend less than they remember spending in previ-
ous weeks. This budget fallacy appears to stem, at least in part, from financial
goals at the time of prediction. Individuals who experienced a stronger

desire to minimize future spending predicted they would be able to reduce
their future spending. However, these goals were not related to subsequent
spending levels, and thus contributed to the prediction bias.
A phenomenon-based approach to research invites the use of multiple
theoretical approaches, and innovation comes not only through traditional
theory-building but also through theoretical integration and application, as
we demonstrated in our extensions to Kahneman and Tversky’s classic
model of the planning fallacy. To understand how people maintain hope
for the future while making sense of the past has required us to integrate a
number of social psychological theories with the inside–outside model of
prediction first proposed by Kahneman and Tversky 40 years ago. Our
extended inside–outside model adds explanatory principles from theories of
motivated reasoning, attribution, visual perspective, temporal construal,
group deliberation, goal-setting, and affective forecasting to the classic
model. The resulting extended model preserves the essential focus on
singular versus distributional reasoning while elaborating on the social
contexts and settings that control not only the relative weight of singular
versus distributional information, but also the content of the singular sce-
nario and its translation into prediction and action. The resulting ‘‘inside
story’’ of the planning fallacy is not simple, but it is coherent and compre-
hensive. As a result, this integrated extended account provides many entry
points for intervention and control, both for use in applied settings and for
individuals hoping to improve their daily planning and prediction.
ACKNOWLEDGMENTS
This program of research was supported by grants from the Social Sciences and Humanities
Research Council of Canada.
REFERENCES
Adams, D., Carwardine, M. (1991). Last chance to see. New York, NY: Harmony Books.
Argote, L., Seabright, M. A., Dyer, L. (1986). Individual versus group use of base-rate and
individuating information. Organizational Behavior and Human Decision Processes, 38,
65–75.
Argote, L., Devadas, R., Melone, N. (1990). The base-rate fallacy: Contrasting processes
and outcomes of group and individual judgment. Organizational Behavior and Human
Decision Processes, 46, 296–310.
Ariely, D., Wertenbroch, K. (2002). Procrastination, deadlines, and performance: Self-
control by precommitment. Psychological Science, 13, 219–224.
Arkes, H. R., Faust, D., Guilmette, T. J., Hart, K. (1988). Eliminating the hindsight bias.
Journal of Applied Psychology, 73, 305–307.
Armor, D. A., Sackett, A. M. (2006). Accuracy, error, and bias in predictions for real
versus hypothetical events. Journal of Personality and Social Psychology, 91, 583–600.

Armor, D. A., Taylor, S. E. (2003). The effects of mindset on behavior: Self-regulation in
deliberative and implemental frames of mind. Personality and Social Psychology Bulletin, 29,
86–95.
Armstrong, J. S. (1985). Long-range forecasting: From crystal ball to computer. New York: Wiley.
Atance, C. M., O’Neill, D. K. (2001). Episodic future thinking. Trends in Cognitive
Sciences, 12, 533–539.
Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change. Psycho-
logical Review, 84, 191–215.
Brooks, F. P. Jr. (1975). The mythical man-month: Essays on software engineering. Reading, MA:
Addison-Wesley.
Brunnermeier, M. K., Papakonstantinou, F., Parker, J. A. (2008). An economic model of the
planning fallacy. Unpublished manuscript, Princeton University.
Buehler, R., Griffin, D. (2003). Planning, personality, and prediction: The role of future
focus in optimistic time predictions. Organizational Behavior and Human Decision Processes,
92, 80–90.
Buehler, R., Griffin, D. (2009). Motivated prediction in observers. Unpublished data.
Buehler, R., McFarland, C. (2001). Intensity bias in affective forecasting: The role of
temporal focus. Personality and Social Psychology Bulletin, 27, 1480–1493.
Buehler, R., Griffin, D., MacDonald, H. (1997). The role of motivated reasoning in
optimistic time predictions. Personality and Social Psychology Bulletin, 23, 238–247.
Buehler, R., Griffin, D., Ross, M. (1994). Exploring the ‘‘planning fallacy’’: Why people
underestimate their task completion times. Journal of Personality and Social Psychology, 67,
366–381.
Buehler, R., Griffin, D., Ross, M. (1995). It’s about time: Optimistic predictions in work
and love. In W. Stroebe M. Hewstone (Eds.), European review of social psychology
(Vol. 6, pp. 1–32). Chichester: Wiley.
Buehler, R., Griffin, D., Ross, M. (2002). Inside the planning fallacy: The causes and
consequences of optimistic time predictions. In T. Gilovich, D. Griffin, D. Kahneman
(Eds.), Heuristics and biases: The psychology of intuitive judgment (pp. 250–270). Cambridge:
Cambridge University Press.
Buehler, R., Messervey, D., Griffin, D. (2005). Collaborative planning and prediction:
Does group discussion affect optimistic biases in time estimation? Organizational Behavior
and Human Decision Processes, 97, 47–63.
Buehler, R., Otsubo, Y., Heine, S., Lehman, D., Griffin, D. (2003). Culture and optimism:
The planning fallacy in Japan and North America. Unpublished manuscript.
Buehler, R., Peetz, J., Griffin, D. (2010a). Finishing on time: When do predictions
influence actual completion times? Organizational Behaviour and Human Decision Processes,
111, 23–32.
Buehler, R., Griffin, D., Lam, K. C. H., Deslaurier, J. (2010b). Perspectives on prediction:
Does third-person imagery improve task completion estimates? Manuscript submitted for
publication.
Burt, C. D. B., Kemp, S. (1994). Construction of activity duration and time management
potential. Applied Cognitive Psychology, 8, 155–168.
Byram, S. J. (1997). Cognitive and motivational factors influencing time predictions. Journal
of Experimental Psychology: Applied, 3, 216–239.
Carver, C. S., Scheier, M. F. (1998). On the self-regulation of behavior. New York: Cam-
bridge University Press.
Cervone, D., Peake, P. K. (1986). Anchoring, efficacy, and action: The influence of
judgmental heuristics on self-efficacy judgments and behavior. Journal of Personality and
Social Psychology, 50, 492–501.
Connolly, T., Dean, D. (1997). Decomposed versus holistic estimates of effort required
for software writing tasks. Management Science, 43, 1029–1045.

Dawes, R. M. (1988). Rational choice in an uncertain world. New York: Harcourt Brace
Jovanovich.
Deslauriers, J. (2002). Should we plan for upcoming tasks? Effects of planning on prediction accuracy
for simple and difficult assignments. Honors Thesis, Wilfrid Laurier University.
Dougherty, M. R. P., Gettys, C. F., Thomas, R. P. (1997). The role of mental simulation
in judgments of likelihood. Organizational Behavior and Human Decision Processes, 70,
135–148.
Dunning, D. (2007). Prediction: The inside view. In A. W. Kruglanski E. T. Higgins
(Eds.), Social psychology: Handbook of basic principles (2nd ed. pp. 69–90). New York:
Guilford Press.
Engerman, S. L., Sokoloff, K. L. (2004). Digging the dirt at public expense: Governance in the
building of the Erie Canal and other public works. NBER Working Paper No. w10965.
Eyal, T., Liberman, N., Trope, Y., Walther, E. (2004). The pros and cons of temporally
near and distant action. Journal of Personality and Social Psychology, 86, 781–795.
Festinger, L. (1957). A theory of cognitive dissonance. Stanford: Stanford University Press.
Fischhoff, B., Slovic, P., Lichtenstein, S. (1978). Fault trees: Sensitivity of estimated failure
probabilities to problem representation. Journal of Experimental Psychology: Human Percep-
tion and Performance, 4, 33–344.
Flyvbjerg, B. (2008). Curbing optimism bias and strategic misrepresentation in planning:
Reference class forecasting in practice. European Planning Studies, 16, 3–21.
Flyvbjerg, B., Bruzelius, N., Rothengatter, W. (2003). Megaprojects and risk: An anatomy of
ambition. United Kingdom: Cambridge University Press.
Flyvbjerg, B., Garbuio, M., Lovallo, D. (2009). Delusion and deception in large infra-
structure projects: Two models for explaining and preventing executive disaster. Cali-
fornia Management Review, 51, 170–193.
Flyvbjerg, B., Holm, M. K. S., Buhl, S. L. (2004). What causes cost overrun in transport
infrastructure projects? Transport Reviews, 24, 3–18.
Flyvbjerg, B., Holm, M. K. S., Buhl, S. L. (2005). How (in)accurate are demand forecasts
in public works projects? Journal of the American Planning Association, 71, 131–146.
Forsyth, D. K., Burt, C. D. B. (2008). Allocating time to future tasks: The effect of task
segmentation on planning fallacy bias. Memory and Cognition, 36, 791–798.
Francis-Smythe, J. A., Robertson, I. T. (1999). On the relationship between time
management and time estimation. British Journal of Psychology, 90, 333–347.
Frank, M. G., Gilovich, T. (1989). The effect of memory perspective on retrospective
causal attributions. Journal of Personality and Social Personality, 57, 399–403.
Gilbert, D. T., Malone, P. S. (1995). The correspondence bias. Psychological Bulletin, 117,
21–38.
Gilovich, T., Kerr, M., Medvec, M. H. (1993). Effect of temporal perspective on
subjective confidence. Journal of Personality and Social Psychology, 64, 552–560.
Gollwitzer, P. M. (1999). Implementation intentions: Strong effects of simple plans. American
Psychologist, 54, 493–503.
Gollwitzer, P. M., Brandstätter, V. (1997). Implementation intentions and effective goal
pursuit. Journal of Personality and Social Psychology, 73, 186–199.
Gregory, W. L., Duran, A. (2001). Scenarios and acceptance of forecasts. In
J. S. Armstrong (Ed.), Principles of forecasting: A handbook for researchers and practitioners
(pp. 519–540). Boston: Kluwer Academic Publishers.
Griffin, D., Brenner, L. (2004). Perspectives on probability judgment calibration. In
D. Koehler N. Harvey (Eds.), Blackwell Handbook of Judgment and Decision Making
(pp. 177–199). Malden: Blackwell Publishing.
Griffin, D., Buehler, R. (1999). Frequency, probability, and prediction: Easy solutions to
cognitive illusions? Cognitive Psychology, 38, 48–78.

Griffin, D., Buehler, R. (2005). Biases and fallacies, memories and predictions: Comment
on Roy, Christenfeld, and Mckenzie (2005). Psychological Bulletin, 131, 757–760.
Griffin, D., Dunning, D., Ross, L. (1990). The role of construal processes in overconfi-
dent predictions about the self and others. Journal of Personality and Social Psychology, 59,
1128–1139.
Guinote, A. (2007a). Power and goal pursuit. Personality and Social Psychology Bulletin, 33,
1076–1087.
Guinote, A. (2007b). Power affects basic cognition: Increased attentional inhibition and
flexibility. Journal of Experimental Social Psychology, 43, 685–697.
Heath, C., Gonzalez, R. (1995). Interaction with others increases decision confidence but
not decision quality: Evidence against information collection views of interactive deci-
sion making. Organizational Behavior and Human Decision Processes, 61, 305–326.
Heine, S. J., Kitayama, S., Lehman, D. R. (2001). Cultural differences in self-evaluation:
Japanese readily accept negative self-relevant information. Journal of Cross-Cultural
Psychology, 32, 434–443.
Heine, S. J., Takata, T., Lehman, D. R. (2000). Beyond self-presentation: Evidence for
self-criticism among Japanese. Personality and Social Psychology Bulletin, 26, 71–78.
Hinds, P. J. (1999). The curse of expertise: The effects of expertise and debiasing methods on
predictions of novice performance. Journal of Experimental Psychology: Applied, 5,
205–221.
Hinsz, V. B., Kalnbach, L. R., Lorentz, N. R. (1997a). Using judgmental anchors to
establish challenging self-set goals without jeopardizing commitment. Organizational
Behavior and Human Decision Processes, 71, 287–308.
Hinsz, V. B., Tindale, R. S., Vollrath, D. A. (1997b). The emerging conceptualization of
groups as information processors. Psychological Bulletin, 121, 43–64.
Hoch, S. J. (1985). Counterfactual reasoning and accuracy in predicting personal events.
Journal of Experimental Psychology. Learning, Memory, and Cognition, 11, 719–731.
Josephs, R. A., Hahn, E. D. (1995). Bias and accuracy in estimates of task duration.
Organizational Behaviour and Human Decision Processes, 61, 202–215.
Kahneman, D., Lovallo, D. (1993). Timid choices and bold forecasts: A cognitive
perspective on risk taking. Management Science, 39, 17–31.
Kahneman, D., Tversky, A. (1979). Intuitive prediction: Biases and corrective procedures.
TIMS Studies in Management Science, 12, 313–327.
Kahneman, D., Tversky, A. (1982a). Intuitive prediction: Biases and corrective proce-
dures. In D. Kahneman, P. Slovic, A. Tversky (Eds.), Judgment under uncertainty:
Heuristics and biases (pp. 414–421). Cambridge: Cambridge University Press.
Kahneman, D., Tversky, A. (1982b). Variants of uncertainty. Cognition, 11, 143–157.
Kahneman, D., Tversky, A. (1982c). The simulation heuristic. In D. Kahneman,
P. Slovic, A. Tversky (Eds.), Judgment under uncertainty: Heuristics and biases
(pp. 201–210). Cambridge: Cambridge University Press.
Kerr, N. L., Tindale, R. S. (2004). Group performance and decision making. Annual
Review of Psychology, 55, 623–655.
Koehler, D. (1991). Explanation, imagination, and confidence in judgment. Psychological
Bulletin, 110, 499–519.
Koehler, D., Poon, C. S. K. (2006). Self-predictions overweight strength of current
intentions. Journal of Experimental Social Psychology, 42, 517–524.
König, C. J. (2005). Anchors distort estimates of expected duration. Psychological Reports, 96,
253–256.
Koole, S., Van’t Spijker, M. (2000). Overcoming the planning fallacy through willpower:
Effects of implementation intentions on actual and predicted task-completion times.
European Journal of Social Psychology, 30, 873–888.

Krizan, Z., Windschitl, P. D. (2007). The influence of outcome desirability on optimism.
Psychological Bulletin, 133, 95–121.
Kross, E., Ayduk, O., Mischel, W. (2005). When asking ‘‘why’’ doesn’t hurt: Distinguish-
ing rumination from reflective processing of negative emotions. Psychological Science, 16,
709–715.
Kruger, J., Evans, M. (2004). If you don’t want to be late, enumerate: Unpacking reduces
the planning fallacy. Journal of Experimental Social Psychology, 40, 586–598.
Kunda, Z. (1990). The case for motivated reasoning. Psychological Bulletin, 108, 480–498.
Ledgerwood, A., Trope, Y., Liberman, N. (2010). Flexibility and consistency in evalua-
tive responding: The function of construal level. In M. P. Zanna J. M. Olson (Eds.),
Advances in Experimental Social Psychology. Maryland Heights, MO: Elsevier.
Libby, L. K., Eibach, R. P. (2002). Looking back in time: Self-concept change and visual
perspective in autobiographical memory. Journal of Personality and Social Psychology, 82,
167–179.
Libby, L. K., Eibach, R. P. (2009). Seeing the links among the personal past, present, and
future: How imagery perspective in mental simulation functions in defining the tempo-
rally extended self. In K. D. Markman, W. M. P. Klein, J. A. Suhr (Eds.), Handbook of
imagination and mental simulation (pp. 359–372). New York: Psychology Press.
Liberman, N., Trope, Y. (1998). The role of feasibility and desirability considerations in
near and distant future decisions: A test of temporal construal theory. Journal of Personality
and Social Psychology, 75, 5–18.
Liberman, N., Sagristano, M. D., Trope, Y. (2002). The effect of temporal distance on
level of mental construal. Journal of Experimental Social Psychology, 38, 523–534.
Liberman, N., Trope, Y., Stephan, E. (2007). Psychological distance. In A. W. Kruglanski
E. T. Higgins (Eds.), Social psychology: Handbook of basic principles (Vol. 2,
pp. 353–383). New York: Guilford Press.
Lichtenstein, S., Fischhoff, B., Phillips, L. D. (1982). Calibration of probabilities: The state
of the art to 1980. In D. Kahneman A. Tversky (Eds.), Judgment under uncertainty:
Heuristics and biases. Cambridge, MA, US: MIT Press.
Locke, E. A., Latham, G. P. (1990). A theory of goal setting and task performance. Englewood
Cliffs, NJ: Prentice Hall.
Locke, E. A., Latham, G. P. (2002). Building a practically useful theory of goal setting and
task motivation: A 35-year odyssey. American Psychologist, 57, 705–717.
Lorenz, C., Neisser, U. (1985). Factors of imagery and event recall. Memory and Cognition,
13, 494–500.
Lovallo, D., Kahneman, D. (2003). Delusions of success: How optimism undermines
executives’ decisions. Harvard Business Review, 81, 56–63.
MacDonald, T. K., Ross, M. (1999). Assessing the accuracy of predictions about dating
relationships: How and why do lovers’ predictions differ from those made by observers?
Personality and Social Psychology Bulletin, 25, 1417–1429.
Marks, D. F. (1999). Consciousness, mental imagery, and action. British Journal of Psychology,
90, 567–585.
McIsaac, H. K., Eich, E. (2002). Vantage point in episodic memory. Psychonomic Bulletin
and Review, 9, 409–420.
Miller, D. T., Ross, M. (1975). Self-serving biases in the attribution of causality: Fact or
fiction? Psychological Bulletin, 82, 213–225.
Molokken, K., Jrgensen, M. (2003). A review of surveys on software effort estimation.
International symposium of empirical software engineering. pp. 223–230.
Morwitz, V. G., Fitzsimons, G. J. (2004). The mere-measurement effect: Why does
measuring intentions change actual behavior. Journal of Consumer Psychology, 14, 64–74.
Mulhern, K. (2006). Self-other differences in task completion predictions: The impact of motivation.
Honors Thesis, Wilfrid Laurier University.

Nelson, R. R. (2007). IT project management: Infamous failures, classic mistakes, and best
practices. MIS Quarterly Executives, 6, 67–78.
Newby-Clark, I. R., Ross, M., Buehler, R., Koehler, D., Griffin, D. (2000). People focus
on optimistic and disregard pessimistic scenarios while predicting task completion times.
Journal of Experimental Psychology: Applied, 6, 171–182.
Norris, K. P. (1971). The accuracy of cost and duration estimates in industrial RD. RD
Management, 2, 25–36.
Peetz, J., Buehler, R. (2009). Is there a budget fallacy? The role of savings goals in the
prediction of personal spending. Personality and Social Psychology Bulletin, 35, 230–242.
Peetz, J., Buehler, R., Wilson, A. E. (2010). Planning for the near and distant future: How
does temporal distance affect task completion predictions? Journal of Experimental Social
Psychology. (in press).
Peterson, C., Semmel, A., von Baeyer, C., Abramson, L. Y., Metalski, G. I.,
Seligman, M. E. P. (1982). The attributional style questionnaire. Cognitive Therapy and
Research, 6, 287–299.
Pezzo, M. V., Litman, J. A., Pezzo, S. P. (2006a). On the distinction between yuppies and
hippies: Individual differences in prediction biases for planning future tasks. Personality
and Individual Differences, 41, 1359–1371.
Pezzo, S. P., Pezzo, M. V., Stone, E. R. (2006b). The social implications of planning:
How public predictions bias future plans. Journal of Experimental Social Psychology, 42,
221–227.
Pronin, E., Ross, L. (2006). Temporal differences in trait self-ascription: When the self is
seen as an other. Journal of Personality and Social Psychology, 90, 197–209.
Robinson, J. A., Swanson, J. A. (1993). Field and observer modes of remembering.
Memory, 1, 169–184.
Roy, M. M., Christenfeld, N. J. S. (2007). Bias in memory predicts bias in estimation of
future task duration. Memory Cognition, 35, 557–564.
Roy, M. M., Christenfeld, N. J. S. (2008). Effect of task length on remembered and
predicted duration. Psychonomic Bulletin and Review, 15, 202–207.
Roy, M. M., Christenfeld, N. J. S., McKenzie, C. R. M. (2005). Underestimating the
duration of future events: Memory incorrectly used or memory bias? Psychological Bulletin,
131, 738–756.
Roy, M. M., Mitten, S. T., Christenfeld, N. J. S. (2008). Correcting memory improves
accuracy of predicted task duration. Journal of Experimental Social Psychology: Applied, 14,
266–275.
Sanna, L. J., Parks, C. J., Chang, E. C., Carter, S. E. (2005). The hourglass is half full or
half empty: Temporal framing and the group planning fallacy. Group Dynamics: Theory,
Research, and Practice, 9, 173–188.
Sanna, L. J., Schwarz, N. (2004). Integrating temporal biases: The interplay of focal
thoughts and accessibility experiences. Psychological Science, 15, 474–481.
Savitsky, K., Medvec, V. H., Charlton, A. E., Gilovich, T. (1998). What, me worry?:
Arousal, misattribution, and the effect of temporal distance on confidence. Personality and
Social Psychology Bulletin, 24, 529–536.
Shaklee, H., Fischhoff, B. (1982). Strategies of information search in causal analysis.
Memory and Cognition, 10, 520–530.
Shaw, G. B. (1903/1948). Man and superman; A comedy and a philosophy. Cambridge:
The University Press.
Shepperd, J. A., Ouellette, J. A., Fernandez, J. K. (1996). Abandoning unrealistic
optimism: Performance estimates and the temporal proximity of self-relevant feedback.
Journal of Personality and Social Psychology, 70, 844–855.
Sherman, S. J. (1980). On the self-erasing nature of errors of prediction. Journal of Personality
and Social Psychology, 39, 211–221.

Shrauger, J. S., Ram, D., Greninger, S. A., Mariano, E. (1996). Accuracy of self-
predictions versus judgments by knowledgeable others. Personality and Social Psychology
Bulletin, 22, 1229–1243.
Silk, T. (2010). Getting started is half the battle: The influence of deadlines and consumer
effort on rebate redemption. Manuscript submitted for publication: cited with
permission.
Taylor, S. E., Brown, J. D. (1988). Illusion and well-being: A social psychological
perspective on mental-health. Psychological Bulletin, 103, 193–210.
Taylor, S. E., Pham, L. B., Rivkin, I. D., Armor, D. A. (1998). Harnessing the imagina-
tion: Mental simulation, self-regulation, and coping. American Psychologist, 53, 429–439.
The Standish Group (1995). The CHAOS report into project failure. Boston, MA: Standish
Group International Inc.
The Standish Group (2009). CHAOS summary 2009. Boston, MA: Standish Group Inter-
national Inc.
Thomas, K. E., Handley, S. J. (2008). Anchoring in time estimations. Acta Psychologica,
127, 24–29.
Thomas, K. E., Newstead, S. E., Handley, S. J. (2003). Exploring the time prediction
process: The effects of task experience and complexity on prediction accuracy. Applied
Cognitive Psychology, 17, 655–673.
Tindale, R. S. (1993). Decision errors made by individuals and groups. In N. J. Castellan
(Ed.), Individual and group decision making: Current issues (pp. 109–124). Hillsdale, NJ:
Erlbaum.
Tindale, R. S., Smith, C. M., Thomas, L. S., Filkins, J., Sheffey, S. (1996). Shared
representations and asymmetric social influence processes in small groups. In E. Witte
J. Davis (Eds.), Understanding group behaviour: Consensual action by small groups (Vol. 1,
pp. 81–103). Mahwah, NJ: Erlbaum.
Trope, Y., Liberman, N. (2003). Temporal construal. Psychological Review, 110, 403–421.
Tversky, A., Shafir, E. (1992). Choice under conflict. The dynamics of deferred decision.
Psychological Science, 3, 358–361.
Van Oorschot, K. E., Bertrand, J. W. M., Rutte, C. G. (2002). An empirical study into the
causes of lateness of new product development projects. University of Eindhoven Working Paper
(WP 100).
Vinokur, A., Burnstein, E. (1974). Effects of partially shared persuasive arguments on
group induced shifts: A group-problem-solving approach. Journal of Personality and Social
Weick, M., Guinote, A. (2010). How long will it take? Power biases time predictions.
Journal of Experimental Social Psychology, 46, 595–604.
Wilkes, A., Norris, K. P. (1972). Estimate accuracy and causes of delay in an engineering
research laboratory. RD Management, 3, 35–36.
Wright, E. F., Luus, C. A. E., Christie, S. D. (1990). Does group discussion facilitate the
use of consensus information in making causal attributions? Journal of Personality and Social
Wright, E. F., Wells, G. L. (1985). Does group discussion attenuate the dispositional bias?
Journal of Applied Social Psychology, 15, 531–546.
Yates, J. F. (1990). Judgment and decision making. Englewood Cliffs, NJ: Prentice Hall.
Zauberman, G., Lynch, J. G. Jr. (2005). Resource slack and propensity to discount delayed
investments of time versus money. Journal of Experimental Psychology: General, 134, 23–37.
Zhang, Y., Fishbach, A. (2010). Counteracting obstacles with optimistic predictions.
Journal of Experimental Psychology: General, 139, 16–31.

The Planning Fallacy Cognitive.pdf

More Related Content

Similar to The Planning Fallacy Cognitive.pdf

More from JoshuaLau29

Recently uploaded

The Planning Fallacy Cognitive.pdf